rust-lightning

Prune inbound HTLC onions once forwarded

We store inbound committed HTLCs' onions in Channels for use in reconstructing
the pending HTLC set on ChannelManager read. If an HTLC has been forwarded to
the outbound edge, we no longer need to persist the inbound edge's onion and
can prune it here.

Mark legacy pre-0.3 inbound htlcs on persist

In 0.3+, we are taking steps to remove the requirement of regularly persisting
the ChannelManager and instead rebuild the set of HTLC forwards (and the
manager generally) from Channel{Monitor} data.

We previously merged support for reconstructing the
ChannelManager::decode_update_add_htlcs map from channel data, using a new
HTLC onion field that will be present for inbound HTLCs received on 0.3+ only.

However, we now want to add support for pruning this field once it's no longer
needed so it doesn't get persisted every time the manager gets persisted. At
the same time, in a future LDK version we need to detect whether the field was
ever present to begin with to prevent upgrading with legacy HTLCs present.

We accomplish both by converting the plain update_add option that was
previously serialized to an enum that can indicate whether the HTLC is from
0.2- versus 0.3+-with-onion-pruned (a variant for the latter is added in the
next commit).

Actual pruning of the new update_add field is added in the next commit.

Trivial: document some fields on MonitorRestoreUpdates

Simplify channel_closed check on manager read

This cleanup falls out of the changes made in the previous commit. Separated
out here for reviewability.

De-dup decode_htlcs from monitor only if channel is closed

We recently added support for reconstructing
ChannelManager::decode_update_add_htlcs on startup, using data present in the
Channels. However, we failed to prune HTLCs from this rebuilt map if a given
inbound HTLC was already forwarded to the outbound edge and in the outbound
holding cell (this bug could've caused us to double-forward HTLCs, fortunately
it never shipped).

As part of fixing this bug, we clean up the overall pruning approach by:
1. If the Channel is open, then it is the source of truth for what HTLCs are
outbound+pending (including pending in the holding cell)
2. If the Channel is closed, then the corresponding ChannelMonitor is the
source of truth for what HTLCs are outbound+pending

Previously, we would only consider the monitor's pending HTLCs, which ignored
holding cell HTLCs.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

Don't double-forward inbounds resolved in holding cell

We recently added support for reconstructing
ChannelManager::decode_update_add_htlcs on startup, using data present in the
Channels. However, we failed to prune HTLCs from this rebuilt map if a given
HTLC was already forwarded+removed from the outbound edge and resolved in the
inbound edge's holding cell.

Here we fix this bug that would have caused us to
double-forward inbound HTLC forwards, which fortunately was not shipped.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

fix: test_payment_path_scoring failing

Add `counterparty_node_id` to `TransactionType` variants

This adds `counterparty_node_id` tracking to all `TransactionType`
variants, enabling downstream users to identify the channel counterparty
associated with each broadcast transaction.

For single-channel variants (`CooperativeClose`, `UnilateralClose`,
`AnchorBump`, `Claim`, `Splice`), the counterparty is stored directly.

For multi-channel variants (`Funding`, `Sweep`), the type now uses
`Vec<(PublicKey, ChannelId)>` to pair each channel with its counterparty.

The `OnchainTxHandler` now stores `counterparty_node_id` and provides a
`set_counterparty_node_id` method for initialization during
deserialization of older data.

Co-Authored-By: HAL 9000
Signed-off-by: Elias Rohrer <dev@tnull.de>

Add `counterparty_node_id` to `TrackedSpendableOutput`

This adds an optional `counterparty_node_id` field to
`TrackedSpendableOutput` and updates the `track_spendable_outputs`
method signatures on both `OutputSweeper` and `OutputSweeperSync` to
accept this new parameter.

The field uses TLV type 3 (odd) for backwards compatibility. When
reading outputs tracked with LDK 0.2 and prior, this field will be
`None`.

Co-Authored-By: HAL 9000
Signed-off-by: Elias Rohrer <dev@tnull.de>

Add `counterparty_node_id` to `Event::SpendableOutputs`

This adds an optional `counterparty_node_id` field to the
`SpendableOutputs` event, providing users with information about which
channel counterparty the spendable outputs belong to.

The field uses TLV type 3 (odd) for backwards compatibility, meaning
older versions will safely ignore it during deserialization. When
reading events serialized by older LDK versions, this field will be
`None`.

Co-Authored-By: HAL 9000
Signed-off-by: Elias Rohrer <dev@tnull.de>

fuzz: assert no stuck payments in chanmon_consistency

After settling all state at 0xff, verify that pending_payments is empty
for all nodes. If payments remain stuck indefinitely, that indicates a
bug in payment resolution.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

fuzz: add MPP payment support to chanmon_consistency

Add multi-path payment (MPP) fuzzing commands that split payments across
multiple channels:

- send_mpp_payment: direct MPP from source to dest using multiple channels
- send_mpp_hop_payment: MPP via intermediate node with multiple channels
on either or both hops

New fuzz commands:
- 0x70: direct MPP 0->1 (uses all 3 A-B channels)
- 0x71: MPP 0->1->2, multi channels on first hop (A-B)
- 0x72: MPP 0->1->2, multi channels on both hops (A-B and B-C)
- 0x73: MPP 0->1->2, multi channels on second hop (B-C)
- 0x74: single-channel MPP 0->1 (all parts over one channel)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

fuzz: expand chanmon_consistency to 6 channels (3 per peer pair)

This expands the channel monitor consistency fuzz test from 2 channels
to 6 channels (3 between A-B and 3 between B-C), enabling future MPP
payment testing.

Changes:
- Extract `connect_peers!` macro from `make_channel!` to avoid duplicate
peer connections
- Create channel arrays `chan_ab_ids[3]` and `chan_bc_ids[3]`
- Store SCIDs in `chan_ab_scids[3]` and `chan_bc_scids[3]`
- Use funding transaction versions 1-6 to avoid txid collisions under
fuzz hashing (which XORs all bytes to a single byte, causing versions
0-5 to collide between A-B and B-C channel pairs)
- Update `test_return!` assertions to expect 3/6/3 channels

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

Let Claude run per-crate/feature tests

Previously, we instructed Claude to always run `./ci/ci-tests.sh` which
doesn't seem to work well in practice, as he seems to also employ some
kind of timeout, potentially leading to him getting stuck.

Here we update `CLAUDE.md` accordingly, having him only run `cargo test`

Hold in-flight monitor updates until background event processing

We previously assumed background events would eventually be processed
prior to another `ChannelManager` write, so we would immediately remove
all in-flight monitor updates that completed since the last
`ChannelManager` serialization. This isn't always the case, so we now
keep them all around until we're ready to handle them, i.e., when
`process_background_events` is called.

This was discovered while fuzzing `chanmon_consistency_target` on the
main branch with some changes that allow it to connect blocks. It was
triggered by reloading the `ChannelManager` after a monitor update
completion for an outgoing HTLC, calling
`ChannelManager::best_block_updated`, and reloading the `ChannelManager`
once again. A test is included that provides a minimal reproduction of
this case.

Rustfmt ChannelManager::process_background_events

Make get_latest_mon_update_id a helper on TestChainMonitor

Rename test default channel config

Now that anchor channels are the default,
rename `test_default_anchors_channel_config` to
`test_default_channel_config` and the
previous default to legacy.

Default to anchor channels

Set `negotiate_anchors_zero_fee_htlc_tx` default
to true.

Remove automatic channel acceptance

Removes the `manually_accept_inbound_channels` config
option. In upcoming commit we will default to anchor
channels which requires users checking if they have
enough onchain funds to cover fees in case of a force
close. Hence, we move to always require users to
manually accept inbound channels.

Fix off-by-one for unfunded channel peers

When accepting channels manually, it would
fail if the # of peers without funded channels
was == `MAX_UNFUNDED_CHANNEL_PEERS`, however,
it should fail if the # of peers > `MAX_UNFUNDED_CHANNEL_PEERS`.

Drop `SplicePrototype` feature

Now that `SplicePrototype` and `SpliceProduction` share the same
feature bit, there's not really any reason to have the
`SplicePrototype` feature at all. Instead, we drop it, leaving only
a `Splice` feature.

Switch `SplicePrototype` feature flag to the prod feature bit

When we shipped 0.2 we used the feature bit 155 to signal splicing,
in line with what eclair was using. However, eclair was actually
using that bit to signal splicing on a previous design which is
incompatible with the current spec.

The result of this was that eclair nodes may attempt to splice
using their protocol and we'd fail to deserialize their splice
message (resulting in a reconnect, which luckily would clear their
splice attempt and return the connection to normal).

As we really need to get off of their feature bit and there's not
much reason to keep using a non-final-spec bit, we simply redefine
`SplicePrototype` to bit 63 here.

Support async signing of interactive-tx initial commitment signatures

This commit allows for an async signer to immediately return upon a call
to `EcdsaChannelSigner::sign_counterparty_commitment` for the initial
commitment signatures of an interactively funded transaction, such that
they can call back in via `ChannelManager::signer_unblocked` once the
signatures are ready. This is done for both splices and dual-funded
channels, though note that the latter still require more work to be
integrated. Since `tx_signatures` must be sent only after exchanging
`commitment_signed`, we make sure to hold them back if they're ready to
be sent until our `commitment_signed` is also ready.

fuzz: fix ChaCha20 encrypt_single_block to preserve data

The fuzzing ChaCha20 implementation's encrypt_single_block was not
copying src to dest, causing encrypted data to be lost (dest remained
zeros). This broke payment flows where metadata is encrypted into
payment_secret - the receiver would decrypt zeros and detect the wrong
payment method (LdkPaymentHash instead of UserPaymentHash), causing
payments to fail with "mismatching preimage".

Fix by making encrypt_single_block copy src to dest (identity
encryption), matching the behavior of the process() method.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>