Since it turned out borking channels on every received error could cause
us to bork channels in case of a sync error with C-lightning, we revert
this for now.
This commit changes the logic when garbage collecting forwarding
packages such that they are removed once when the function is called,
and then again upon subsequent ticks. This allows us to bump the
peer timer to 1 hour to limit the number of db transactions happening
in lnd. The forwarding packages need to be removed initially as
otherwise a flappy node will never have them garbage collected.
As part of the preparation to the switch interceptor feature, this
function is changed to return error instead of error channel that
is closed automatically.
Returning an error channel has become complex to maintain and
implement when adding more asynchronous flows to the switch.
The change doesn't affect the current behavior which logs the
errors as before.
This commit extends the link with a new synchronous delivery point for
local UpdateAddHTLC messages. The switch method SendHTLC is updated to
use this delivery point and thereby becomes a synchronous call.
For MPP payments, synchronous hand-off is important. Otherwise the next
pathfinding round could start without the channel balance updated yet.
This commit adds a PendingCommitTicker to the link config, which allows
us to control how quickly we fail the link if the commitment dance
stalls. Now that the mailbox has the ability to cancel packets, when the
link fails it will reset the mailbox packets on exit, forcing a
reevaluation of the HTLCs against their mailbox expiries.
Now that packet failure is handled by the mailbox, we can now enforce
a delivery deadline and fail the packet if it the deadilne is exceeded.
This gives senders quicker feedback about tried routes, and allows them
to try alternative paths to the destination in the meantime.
This commit moves the current logic for sending failures out of the link
and into the mailbox in preparation for our failing delayed htlcs. We do
so because the mailbox may need to fail packets while the link is
offline, and needs to be able to complete the task without member
methods on the link.
This commit removes the overflowQueue from the link. We do so in order
to promote better UX for senders, so that HTLCs are failed faster when
the commitment is full. This gives the sender the opportunity to try
another, more open path, rather than perceive the HTLC as being stuck.
At the same time, we remove the total number of active goroutines in lnd
by a factor of N where N is the number of active channels.
The linter complains about not checking the return value from
WipeChannel in certain places. Instead of checking we simply remove the
returned error because the in-memory modifications cannot fail.
This commit modifies updateCommitTx to error with ErrLinkShuttingDown
when we try to send a ContractUpdate on the htlcUpdates chan and the
link has closed the quit chan. It also changes the order of the call
to ackDownStreamPackets and contract update call for consistency since
the packets should be acknowledged before the link goes down.
This commit adds link failure notifications for failures which occur
on our incoming link. These failures may be receives which we failed or
forwards which we could not parse.
This commit adds notifications for htlcs which are forwarded through
our node. Forwards are notified when the htlc is added on our ougoing
link, settles when we send a settle message to the downstream peer.
If a failure occurs, we check whether it occurred at our node, then
notify a link or forwarding failure accordingly.
Note that this change also adds forward event notifications for sends
which are initiated by our node because the handling code for adding
a htlc which originates from our node is the same as that for handling
forwards. Htlcs for our locally initiated sends have our internal pid
set in the incoming htlcs id field, so we extract this value and notify
with a zero htlc id to be consistent with receives (which have zero
outgoing circuits). Subsequent settles or failures are not noitfied
for local sends in this commit, and will be handled in a follow up.
In this commit, a htlcNotifier interface is added to allow for easy
unit testing. Instances of the HtlcNotifier are added to the server,
switch and link.
Since we want to handle the edge case where paying the HTLC fee would
take the initiator below the reserve, we move the subtraction of the
reserve into availableBalance where this calculation will be performed.
This commit adds LinkErrors with failure details to htlcs which fail on
our incoming link. This change is made with the intention of notifying
detailed htlc failure reasons in sendHTLCError. The FailureDetail
interface is implemented on FailureResolutionResults so that they can
directly be used to enrich LinkErrors. sendHtlcError is updated to
take a LinkError in preparation for the addition of a htlcnotifier
which will notify the detail of the error.
This commit adds a linkError field to track the value of failures
which occur at our node. This field is set when local payments or
multi hop htlcs fail in the switch or on our outgoing link. This
addition is required for the addition of a htlc notifier which will
notify these failures in handleDownstreamPacket.
The passing of link error to failAddPacket removes the need for an
additional error field, because the link error's failure detail will
contain any additional metadata. In the places where the failure detail
does not cover all the metadata that was previously supplied by addr
err, the error is logged before calling failAddPacket so that this
change does not reduce the amount of information we log.
Rename FailureDetail in a separate commit so that a FailureDetail
interface can be introduced in the following commit.
OutgoingFailureOnionDecode is renamed to OutgoingFailureDecodeError
to specifically indicate that we could not decode the wire
failure that our payment experienced.
This commit repalces the htlcResolution struct with an interface.
This interface is implemeted by failure, settle and accept resolution
structs. Only settles and fails are exported because the existing
code that handles htlc resolutions uses a nil resolution to indicate
that a htlc was accepted. The accept resolution is used internally
to report on the resolution result of the accepted htlc, but a nil
resolution is surfaced. Further refactoring of all the functions
that call NotifyExitHopHtlc to handle a htlc accept case (rather than
having a nil check) is required.
Update the ChannelLink interface to specifically
return the LinkError struct. This error implements
the ClearTextError interface, so will be picked
up as a routing realted error by the router.
With LinkErrors implemented, the switch now
returns a LinkError for all failures on our
incoming/outgoing link and ForwardingError when
the failure occurs down the line.
This commit adds a getResolutionFailure function
which returns an appropriate wire failure based
on the outcome of a htlc resolution. It also updates
the MissionControlStore test to ensure that lnd
can handle failures which occur due to mpp timeout.
This commit moves handling of invoice not found
errors into NotifyExitHopHtlc and exposes a
resolution result to the calling functions. The
intention of this change is to make calling
functions as naive of the invoice registry's
mechanics as possible.
When NotifyExitHopHtlc is called and an invoice
is not found, calling functions can take action
based on the HtlcResolution's InvoiceNotFound
outcome rather than having to add a special error
check on every call to handle the error.
This commit adds the resolution result obtained
while updating an invoice in the registry to
htlcResolution. The field can be used by calling
functions to determine the outcome of the
update and act appropriately.
This commit renames HodlEvent to HtlcResolution
to better reflect the fact that the struct is
only used for htlc settles and cancels, and that
it is not specifically used for hodl invoices.
Replace logCommitTick as a way to deal with revocation window exhaustion
by retrying to update the commit tx when the remote revocation is
received.
The rationale is that the revocation window always opens up because of a
revoke message that is received from the other party. It is therefore
not necessary to set a timer for this. The reception of the revoke
message is the trigger to send a new commit sig if necessary.
