This commit adds a listener queue for each peer, that can be used
to queue listeners that will be notified when the targetted peer
eventually comes online.
This commit adds a new primary goroutine to the server struct:
peerBootstrapper. If peer boostrapping isn’t disabled in the config,
this new goroutine will be launched to attempt to establish a set of
initial connections for a new node. The logic is pretty straight
forward: first a set of initial connections is attempted, if after our
first epoch, we don’t have enough connections yet, then we’ll attempt
to query for an additional set. In each iteration, if we haven’t been
successful, then we increase our exponential backoff in order to not
spam any of our bootstrapping sources.
This commit alters the synchronization patterns used in the server
such that the internal state is protected by a single mutex. Overall,
this simplifies the ability to reason about the behavior and
manipulation of the internal state, which has resolved a few of flakes
related to race conditions that were observed before hand.
Invoking DisconnectPeer is now fully synchronous, and waits until
the provided peer's peerTerminationWatcher has exited before
returning. Currently this is done by tracking the watcher using the
peer's WaitGroup, and locking until the peer has shutdown.
The server's API has also been refactored such that all public methods
are safe for concurrent use. Therefore, other subsystems should be
sure to make use of these endpoints to avoid corrupting the internal
state.
This commit adds a new utility method to the server struct itself. This
method will allow callers to query the state of the server in order to
decide if the server has been started or not. This can be useful
elsewhere in the project as we start to decouple the lifetime of
certain sub-systems from others.
This commit updates the main single-funder funding workflow within the
fundingManager (initiated via the rpcserver or by a message from a
connected peer) to fully adhere to the funding protocol outlined in
BOLT-0002.
The major changes are as follows:
* All messages modified to use the new funding messages in BOLT-0002.
* The initiator of a funding workflow no longer decides how many
confirmations must elapse before the channel can be considered open.
* Rather than each side specifying their desired CSV delay, both
sides now specify the CSV delay for the _other_ party.
In previous commits we have intoduced the onion errors. Some of this
errors include lnwire.ChannelUpdate message. In order to change
topology accordingly to the received error, from nodes where failure
have occured, we have to propogate the update to the router subsystem.
In this commit daemon have been changed to set the proper hooks in the
channel link and switch subsystems so that they could send and receive
encrypted onion errors.
This commit overhauls the way that lnd is created with the goal of
ensuring the chain backends are fully synced up before the daemon
itself starts. The rpcServer has been slightly decoupled from the
server itself s.t we can start the rpcServer independently of the
server. This is required as we’ll now wait (unless we’re in simnet
mode) for the chain to fully sync up before we even _start_ any of the
server’s goroutines.
This commit modifies the Peers method on the server such that a caller
is able to query this method without the main serger goroutines
started. This is a small component in a larger change which will let us
start the RPC server independently of the server.
In current commit big shift have been made in direction of unit testable
payments scenarios. Previosly two additional structures have been added
which had been spreaded in the lnd package before, and now we apply
them in the lnd itself:
1. ChannelLink - is an interface which represents the subsystem for
managing the incoming htlc requests, applying the changes to the
channel, and also propagating/forwarding it to htlc switch.
2. Switch - is a central messaging bus for all incoming/outgoing htlc's.
The goal of the switch is forward the incoming/outgoing htlc messages
from one channel to another, and also propagate the settle/fail htlc
messages back to original requester.
With this abtractions the folowing schema becomes nearly complete:
abstraction
^
|
| - - - - - - - - - - - - Lightning - - - - - - - - - - - - -
|
| (Switch) (Switch) (Switch)
| Alice <-- channel link --> Bob <-- channel link --> Carol
|
| - - - - - - - - - - - - - TCP - - - - - - - - - - - - - - -
|
| (Peer) (Peer) (Peer)
| Alice <----- tcp conn --> Bob <---- tcp conn -----> Carol
This commit adds the FeeEstimator interface, which can be used for
future fee calculation implementations. Currently, there is only the
StaticFeeEstimator implementation, which returns the same fee rate for
any transaction.
This commit fixes a bug that was introduced when the connection
handling was re-worked to properly handle the case of concurrent
connections being made. In certain cases after a successful initial
connection, a peer’s stray goroutine would still attempt to establish a
second outbound connection even though a connection had already been
established. This was properly handled by the connecting peer, but not
he receiving peer. This commit adds the additional logic to the
receiving peer to ensure that we properly handle this case.
In this commit waiting proofs array have been replaced with persistant
boltd storage which removes the possibility for the half proof to be
lost during half proof exchange.
This commit modifies the implementation of the new DisconnectPeer RPC
in the following ways:
* all validation has moved from the server to the rpcserver
* rather than iterating over _all_ channels, we now only check the
peer’s channels
* the disconnectPeerMsg now has a public key object
* this allows us to also verify that the user submitted a valid
pub key string
* we now check if a peer was persistent when disconnecting so we can
remove them from the persistent peer map
This commit fixes a prior bug wherein if a user connected to a peer
using the —perm command, then once the peer was disconnected, we
wouldn’t automatically connect to them.
Issue: 139
This commit contains client-side and server-side functionality
for disconnecting peers. rpc-client calls server side method and sends
message with pubKey.
This commit fixes a bug that was introduced when the concurrent
connection handling logic was re-written: if we don’t properly add the
persistent outbound connection to the persistent conn reqs map. The fix
is easy: add the pending conn req to the proper map.
This commit fixes a bug that would possibly result in tens of goroutine
beaching launched in an attempt to persistently connect to a peer. This
bug has been fixed by ensuring that we’ll only launch a new pending
connection attempt if we don’t already have one pending.
The prior methods we employed to handle persistent connections could
result in the following situation: both peers come up, and
_concurrently_ establish connection to each other. With the prior
logic, at this point, both connections would be terminated as each peer
would go to kill the connection of the other peer. In order to resolve
this issue in this commit, we’ve re-written the way we handle
persistent connections.
The eliminate the issue described above, in the case of concurrent peer
connection, we now use a deterministic method to decide _which_
connection should be closed. The following rule governs which
connection should be closed: the connection of the peer with the
“smaller” public key should be closed. With this rule we now avoid the
issue described above.
Additionally, each peer now gains a peerTerminationWatcher which waits
until a peer has been disconnected, and then cleans up all resources
allocated to the peer, notifies relevant sub-systems of its demise, and
finally handles re-connecting to the peer if it's persistent. This
replaces the goroutine that was spawned in the old version of
peer.Disconnect().
This commit re-writes the GetNetworkInfo implenetaiton to use a single
database transaction. We’re now able to do this due to the recent
change in the API for the ChannelGraph struct and it’s related objects.
The recent change allows the passed callback to accept a db
transaction, with this, the callback is now able to issue another
traversal routine _within_ the prior one.
This commit modifies the fundingManager config to use the a SignMesage
function rather than two distinct functions for singing one half the
channel announcement proofs. This change unifies the signing of
messages under a single abstraction: the MessageSigner interface.
This commit eliminates a possible deadlock (or repeated peer connection
failures) that can arise due to the [inbound|outbound]PeerConnected
methods holding the peer mutex too long. We now alleviate this
concurrency issue by calling s.peerConnected in an asynchronous manner.
This is safe as all operations within the method are themselves
goroutine-safe.
