Releasing write locks in between monitor updates
requires storing a set of cloned keys to iterate
over. For efficiency purposes, that set of keys
is an actual set, as opposed to array, which means
that the iteration order may not be consistent.
The test was relying on an event array index to
access the revocation transaction. We change that
to accessing a hash map keyed by the txid, fixing
the test.
Previously, updating block data on a chain monitor
would acquire a write lock on all of its associated
channel monitors and not release it until the loop
completed.
Now, we instead acquire it on each iteration,
fixing #2470.
- Split Score from LockableScore to ScoreLookUp to handle read
operations and ScoreUpdate to handle write operations
- Change all struct that implemented Score to implement ScoreLookUp
and/or ScoreUpdate
- Change Mutex's to RwLocks to allow multiple data readers
- Change LockableScore to Deref in ScorerAccountingForInFlightHtlcs
as we only need to read
- Add ScoreLookUp and ScoreUpdate docs
- Remove reference(&'a) and Sized from Score in ScorerAccountingForInFlightHtlcs
as Score implements Deref
- Split MultiThreadedScoreLock into MultiThreadedScoreLockWrite and MultiThreadedScoreLockRead.
After splitting LockableScore, we split MultiThreadedScoreLock following
the same way, splitting a single score into two srtucts, one for read and
other for write.
MultiThreadedScoreLock is used in c_bindings.
This previously led to a debug panic in the router because we wouldn't account
for the blinded path fee when calculating first_hop<>intro_node hop's available
liquidity and construct an invalid path that forwarded more over said hop than
was actually available.
This also led to us hitting unreachable code, see direct_to_matching_intro_nodes
test description.
The BOLT spec mandates that channels not be announced until they
have at least six confirmations. This is important to enforce not
because we particularly care about any specific DoS concerns, but
because if we do not we may have to handle reorgs of channel
funding transactions which change their SCID or have conflicting
SCIDs.
Because a `UtxoLookup` implementation is likely to need a reference
to the `PeerManager` which contains a reference to the
`P2PGossipSync`, it is likely to be impossible to get a mutable
reference to the `P2PGossipSync` by the time we want to add a
`UtxoLookup` without a ton of boilerplate and trait wrapping.
Instead, we simply place the `UtxoLookup` in a `RwLock`, allowing
us to modify it without a mutable self reference.
The lifetime bounds updates in tests required in this commit are
entirely unclear to me, but do allow tests to continue building, so
somehow make rustc happier.
In LDK, we expect users operating nodes on the public network to
implement the `UtxoSource` interface in order to validate the
gossip they receive from the network.
Sadly, because the DoS attack of flooding a node's gossip store
isn't a common issue, and because we do not provide an
implementation off-the-shelf to make doing so easily, many of our
downstream users do not have a `UtxoSource` implementation.
In order to change that, here we implement an async `UtxoSource`
in the `lightning-block-sync` crate, providing one for users who
sync the chain from Bitcoin Core's RPC or REST interfaces.
In 3f32f60ae7 we exposed the
historical success probability buckets directly, with a long method
doc explaining how to use it. While this is great for logging
exactly what the internal model thinks, its also helpful to let
users know what the internal model thinks the success probability
is directly, allowing them to compare route success probabilities.
Here we do so but only for the historical tracking buckets.
When we attempt to score a channel which has a success probability
very low, we may have a log well above our cut-off of two. For the
liquidity penalties this works great, we bound it by
`NEGATIVE_LOG10_UPPER_BOUND` and `min` the two scores. For the
amount liquidity penalty we didn't do any `min`ing at all.
This fix is to min the log itself first and then reuse the min'd
log in both calculations.
Currently we let an `htlc_amount >= channel_capacity` pass through
from `penalty_msat` to
`calculate_success_probability_times_billion`, but only if its only
marginally bigger (less than 65/64ths). This is fine as
`calculate_success_probability_times_billion` handles bogus values
just fine (it will always return a zero probability in such cases).
However, this is risky, and in fact breaks in the coming commits,
so instead check it before ever calling through to the historical
bucket probability calculations.
Here we implement `WatchtowerPersister`, which provides a test-only
sample implementation of `Persist` similar to how we might imagine a
user to build watchtower-like functionality in the persistence pipeline.
We test that the `WatchtowerPersister` is able to successfully build and
sign a valid justice transaction that sweeps a counterparty's funds if
they broadcast an old commitment.
For watchtowers to be able to build justice transactions for our
counterparty's revoked commitments, they need to be able to find the
revokeable output for them to sweep. Here we cache `to_self_delay` in
`CommitmentTransaction` to allow for finding this output on the struct
directly. We also add a simple helper method to aid in building the
initial spending transaction.
This also adds a unit test for both of these helpers, and
refactors a bit of a previous `CommitmentTransaction` unit test to make
adding these easier.
Upon creating a channel monitor, it is provided with the initial
counterparty commitment transaction info directly before the very first
time it is persisted. Because of this, the very first counterparty
commitment is not seen as an update in the persistence pipeline, and so
our previous changes to the monitor and updates cannot be used to
reconstruct this commitment.
To be able to expose the counterparty's transaction for the very first
commitment, we add a thin wrapper around
`provide_latest_counterparty_commitment_tx`, that stores the necessary
data needed to reconstruct the initial commitment transaction in the
monitor.
Currently, when we receive an HTLC claim from a peer, we first hash
the preimage they gave us before removing the HTLC, then
immediately pass the preimage to the inbound channel and hash the
preimage again before removing the HTLC and sending our peer an
`update_fulfill_htlc`. This second hash is actually only asserted
on, never used in any meaningful way as we have the htlc data
present in the same code.
Here we simply drop this second hash and move it into a
`debug_assert`.
If a user has issues with a payment, the most obvious thing they'll
do is check logs for the payment hash. Thus, we should ensure our
logs that show a payment's lifecycle include the payment hash and
are emitted (a) as soon as LDK learns of the payment, (b) once the
payment goes out to the peer (which is already reasonably covered
in the commitment transaction building logs) and (c) when the
payment ultimately is fulfilled or fails.
Here we improve our logs for both (a) and (c).
This adds the feerate and local and remote output values to this channel
monitor update step so that a monitor can reconstruct the counterparty's
commitment transaction from an update. These commitment transactions
will be exposed to users in the following commits to support third-party
watchtowers in the persistence pipeline.
With only the HTLC outputs currently available in the monitor update, we
can tell how much of the channel balance is in-flight and towards which
side, however it doesn't tell us the amount that resides on either side.
Because of dust, we can't reliably derive the remote value from the
local value and visa versa. Thus, it seems these are the minimum fields
that need to be added.
