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 splits the packet courier internally into two distinct
queues, one for adds and one for settles+fails. This allows us to
prioritize HTLCs that will clear the commitment transaction and make
space for adds. Previously this responsibility was handled by the
overflow queue.
This commit delays the advancement of the pktHead until after the
message has been delivered. This is a prepatory step, as in the future
we may fail to deliver the packet due to a deadline expiring.
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 is less brittle than blacklisting states since it may not be clear
that new states needs to be added. Now that HasChanStatus is fixed for
ChanStatusDefault, we can safely implement this.
This addresses a bug in which a funding reservation wasn't cleaned up
properly if the remote peer didn't support upfront shutdown.
Alternatively, we could just cancel the reservation on error, but
instead we move the check above so that we don't attempt coin selection
in the first place.
We'll do the validation during construction of the runtime so that we
can safely use the AppPreRelease field externally without needing to
normalize it.
This commit removes ChanStatusDefault from the list of
orderedChanStatusFlags since it is not flag. As with the prior commit,
the logic around these flags assumes everything in the list is a flag,
but ChanStatusDefault is not.
It turns out we properly special case that if the channel is in
ChanStatusDefault that we only return ChanStatusDefault. However, if any
of the bits are set we would always report ChanStatusDefault since
status&0 == 0. This fixed simply by removing ChanStatusDefault from the
list since we only need the list to express non-default status flags.
This commit resovles a lingering issue w/in the codebase wrt how the
ChannelStatus flags are defined. Currently ChannelStatus is improperly
used to define a bit field and the individual flags themselves. As a
result, HasChanStatus accepts queries on particular status (combinations
of flags) and individual flags themselves.
This is an issue because the way HasChanStatus computes whether the
channel has a particular status assumes the provided inputs are all
flags (or at least combinations of flags).
However, ChanStatusDefault is simply the absence of any other flag.
Hence, HasChanStatus will always return true when querying for
ChanStatusDefault because status&0 == 0 is always true.
Longer term we should should consider splitting these definitions into
flags and particular states, and change the way construct or operate on
them, but for now I've just special-cased this one value. Fortunately,
we don't query HasChannelStatus w/ ChanStatusDefault anywhere in the
codebase so we dodge a bullet here, but it'd be nice to have some
greater assurances moving forward.
This commit introduces a new test case that asserts all of the witness
size constants currently in the codebase. We also reintroduce the
AcceptedHtlcSuccessWitnessSize and OfferedHtlcTimeoutWitnessSize
constants that were recently removed for the sake of completeness.
In asserting the witnes sizes, there were three uncovered discrepancies:
* OfferedHtlcSuccessWitnessSize overestimated by about 30% because it
included an extra signature in the calculation.
* ToLocalPenaltyWitnessSize was underestimated by one byte, because it
was missing the length byte for the OP_TRUE. This has implications
the watchtower protocol since the client and server are assumed to
share the same weight estimates used for signing. This commit keeps
the current behavior, with the intention of rolling out negotiation
for which weight estimate to use for a given session.
* AcceptedHtlcScriptSize was underestimated by one byte because it was
missing a length byte for the value 32 pushed on the stack when
asserting the preimage's length. This affects all AcceptedHtlc*
witness sizes.
In this commit, we remove the restriction surrounding the largest
invoices that we'll allow a user to create. After #3967 has landed,
users will be able to send in _aggregate_ a payment larger than the
current max HTLC size limit in the network. As a result, we can just
treat that value as the system's MTU, and allow users to request
payments it multiples of that MTU value.
A follow up to this PR at a later time will also allow wumbo _channels_.
However, that requires us to tweak the way we scale CSV values, as post
wumbo, there is no true channel size limit, only the
_local_ limit of a given node. We also need to implement a way for nodes
to signal to other nodes their accepted max channel size.