There is a sleep after channels are opened to ensure the channel is
removed from the set of pending reservations. We can avoid this sleep
and get better guarantees of the channel being opened by deleting the
reservation just before updating channel state to active instead of
after.
This commit decouples the wait for funding transaction confirmations
in the waitForFundingConfirmation function from the announcement of
the channel in the sendFundingLockedAndAnnounceChannel function.
Additionally, the sendFundingLockedAndAnnounceChannel function is
now decoupled into the sendFundingLocked and sendChannelAnnouncement
functions. There is also now a helper function that houses creation
of a lnwire.LightningChannel object, calls to both sendFundingLocked
and sendChannelAnnouncement.
This commit adds a select statement for sending on the fundingManager's
arbiterChan and the peer's newChannels channel. This makes sure we won't
be blocked sending on these channels in case of a shutdown.
The fundingManager will register with the server to get notified
when the targetted peer comes online, in case of a failed send
of the fundingLocked message. This is necessary because if the
peer is not connected yet (or was disconnected while we were
waiting for the funding tx to confirm), we cannot continue the
the opening process before the peer successfully has received
the fundingLocked.
This commit adds a channel barrier on fundingManager startup for
channels where the opening process is not finished. This fixes
a bug where we after restarting the fundingManager would receive
the fundingLocked message, and crash when trying to close the
non-existing barrier.
In case we received a fundingLocked message after our own opening
process was finished, we would crash with the same error. We
therefore check if the channel barrier exists before we try to
close it.
It also adds functionality to fundingManager that makes it
ignore a fundingLocked message it receives for a channel where
this is already received. This is necessary when we in case of
a reconnection resend the fundingLocked since we cannot be sure
the remote has received it.
The fundingmanager tests are also updated to check that the
fundingLocked messages are sent and handled correcly, and also
exercise the scanarios described above.
This commit adds a conditional send on quit to all methods used by
outside sub-systems to the funding manager. This ensures, that in the
case the funding manager is exiting, the caller won’t block
indefinitely.
This commit corrects a minor formatting error when logging the pending
channel ID within the logs. Previously, the logging directives and
parameter could cause the pending chan ID to display in a double-hex
encoded format. We fix this by ensuring that we properly slice the chan
ID before printing it, and also ensure that we use the %x formatting
(which will hex encode the bytes) everywhere.
Fixes#331.
This commit implements some missing functionality as we’ll now properly
validate and generate the various channel flow control constraints
during initial channel funding.
With this commit, we take an additional step towards full spec
compliance as we’ll now properly send over the required channel
reservation, max HTLC’s, and other parameters during the funding flow.
When processing the desired parameters by the remote party, if we think
they’re unreasonable, then we’ll send an Error message and end the
funding flow.
Previously, our old lnwire.Error message used a special code to
indicate different types of errors. With the recent push for spec
compatibility, we removed this field and instead stuffed the error into
the first byte of the error data. This works between lnd nodes, but
with other implementations they may send different errors which use a
different error scheme.
To fix this, we’ll now unconditionally return the error to the end
caller (if one exists).
This commit fixes a prior goroutine leak that could result in a node
having thousands of goroutines, particularly due to many concurrent
channel fundings. We now ensure that for each BlockEpoch client
created, we ensure that the client is cancelled once the creating
grouting exits.
This commit updates the tail end of the funding workflow to properly
include the ChainHash field when crafting ChannelAnnouncements and the
initial ChannelUpdate messages. Additionally, we now properly generate
the proper signatures to match the changes to BOLT0007.
This commit modifies the generation and parsing of errors to abandon
usage of the Code field, and instead use the first byte of the Data
field to store the error codes that we currently use. With this change,
we ensure that our error messages properly match BOLT-0002.
Persists the state of a channel opening process after funding
transaction is confirmed. This tracks the messages sent to
the peer such that the process can be continued in case of a
restart. Also introduces that the receiver side forgets about
channel if funding transaction is not confirmed in 48hrs.
Make the fundingmanager send an updated node announcement
each time it opens a new channel. This is to make sure
our node announcement is propagated in the network, since
peers will ignore our node announcements if we haven't
opened any channels yet.
This commit ensures that we always clean up the resources that are
created when a new instance of a lnwallet.LightningChannel is
instantiated. The is necessary due to the sigPool that’s now present as
an internal goroutine.
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.
This commit fixes a possible panic within the funding manger’s workflow
for pending channels. We now ensure that the local discovery signal
retrieved from the localDiscoverySignals map is always non-nil.
Otherwise, we risk a server panic in the case that a node retransmits
the FundingLocked message after a channel has been fully processed, or
even just sends a FundingLocked message for a non-existent channel.
This commit adds a new bit of synchronization to the funding manager to
ensure that we only process the funding locked message directly _after_
the short channel ID has become available within the channel. This
fixes a possible bug wherein we would receive the funding locked
message, and register the channel with the switch without yet knowing
the short chan ID. This would then cause any HTLC’s routed to the new
channel to fail, as it would be using the incorrect short channel ID.
This commit modifies the funding process to use the short channel ID,
rather than only the opening block height to mark a channel as open
once it has been confirmed. With this change, the short channel ID
information will now be available immediately after the channel has
been confirmed in the chain.
This commit modifies the funding workflow slightly to move the querying
to the fee estimator (for the new channel) into the fundingManager
rather than within the LightningWallet layer. When querying to
FeePerWeight, we now multiply by 1000 to arrive at fee-per-kw.
Additionally, we now also properly thread through the fee-per-kw
offered by the initiator the to the responder of the channel workflow.
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.
In order to make the node's dust limit available to the wallet during
the initial stages of the funding process, add and set a
DustLimit field in the Contribution.
This commit modifies the OpenChannel structure on-disk to also track
that opening height of a channel. This change is being made in order to
make and more light client friendly. A follow up commit will modify
several areas of the codebase to use this new functionality.
This commit fixes a bug that double counts the number of pending
channels in GetInfo. Previously we weren’t yet storing the pending
channels on disk, we are now but comparing both the disk channel and
the channels within memory leads us to double count channels. To fix
this, we now only count the database channels.
Note that this NumPendingChannels method can now be removed as it’s no
longer needed.
This commit modifies the order of operations after a channel has been
detected as fully opened. Rather than waiting until we receive the
FundingLocked message to announce the channel, we now do so immediately
after we detect the channel has been opened.
Additionally, we also patch a possible dead-lock bug by ensuring that
the RevokeAndAck message is always sent _after_ the FundingLocked
message. We do this by only passing the newly created channel too the
relevant sub-systems _after_ we receive the FundingLocked message. This
movement also serves to pave the way for our switch to the spec’s
funding workflow, as once we remove the initial revocation window,
there’ll be no way for us to initiate a state transition until we
receive the FundingLocked message from the remote party.
This commit modifies the way the fundingManager tracks pending funding
workflows internally. Rather than using the old auto-incrementing
64-bit pending channel ID’s, we now use a 32-byte pending channel ID
which is generated using a CSPRG. Additionally, once the final funding
message has been sent, we now de-multiplex the FundingLocked message
according to the new Channel ID’s which replace the old ChannelPoint’s
and are exactly 32-bytes long.
