]>
git.bitcoin.ninja Git - rust-lightning/log
Matt Corallo [Sat, 3 Nov 2018 03:59:59 +0000 (23:59 -0400)]
Merge pull request #241 from TheBlueMatt/2018-10-peer-free-panic
Fix pre-noise outbound peer disconnect panic found by fuzzer
Matt Corallo [Sat, 3 Nov 2018 03:07:54 +0000 (23:07 -0400)]
Merge pull request #233 from TheBlueMatt/2018-10-shutdown-updates
Shutdown Updates
Matt Corallo [Fri, 2 Nov 2018 01:45:59 +0000 (21:45 -0400)]
Test the various shutdown handling updates
Matt Corallo [Thu, 1 Nov 2018 21:17:28 +0000 (17:17 -0400)]
Rebroadcast shutdown on channel_reestablish (and reprocess them)
Matt Corallo [Thu, 1 Nov 2018 21:12:20 +0000 (17:12 -0400)]
Delay closing_signed until update_fee exchanges complete
See https://github.com/lightningnetwork/lightning-rfc/issues/499
for a bit more about the ambiguity we're addressing here.
Also drop holding cell update_fee the same way we drop holding
cell update_add_htlcs when sending shutdown, resolving a bug.
Matt Corallo [Wed, 31 Oct 2018 01:47:56 +0000 (21:47 -0400)]
Fail incoming HTLCs sent after we start shutdown, not the chan
Matt Corallo [Fri, 2 Nov 2018 14:54:25 +0000 (10:54 -0400)]
Avoid double-panic in full_stack_target to assist debugging
Matt Corallo [Fri, 2 Nov 2018 14:50:32 +0000 (10:50 -0400)]
Fix pre-noise outbound peer disconnect panic found by fuzzer
If we make an outbound connection to a peer who we are already
connected to, and the outbound connection fails
pre-noise-completion, we will remove the original peer connection
from our node_id_to_descriptor map.
The fuzzer managed to find this by crashing in Channel's assertions
that we don't do a get_channel_reestablish() when the Channel isn't
already marked disconnected.
Matt Corallo [Fri, 2 Nov 2018 14:45:29 +0000 (10:45 -0400)]
Add logging for (dis)connect in peer_handler/channelmanager
Matt Corallo [Tue, 30 Oct 2018 20:25:38 +0000 (16:25 -0400)]
Send closing_signed when appropriate as pending HTLCs clear
Matt Corallo [Tue, 30 Oct 2018 19:53:34 +0000 (15:53 -0400)]
Always send Shutdown resposnes to Shutdown messages
We always require that any changes to Channel state be committed
immediately (within the same lock) so we should never have
uncommitted changes which would prevent us from sending a Shutdown
response.
Matt Corallo [Tue, 30 Oct 2018 18:39:20 +0000 (14:39 -0400)]
Refactor out first-closing-sent creation
Matt Corallo [Mon, 29 Oct 2018 21:38:16 +0000 (17:38 -0400)]
Use ChannelError in shutdown handling, clean up, impl err actions
Matt Corallo [Tue, 30 Oct 2018 00:21:47 +0000 (20:21 -0400)]
Add warnings reminding users to use SegWit funding inputs
Matt Corallo [Thu, 1 Nov 2018 02:44:16 +0000 (22:44 -0400)]
Merge pull request #235 from TheBlueMatt/2018-10-229-redux
Bug fix using same seed sequence for channel key geneartion
Yuntai Kyong [Sun, 28 Oct 2018 05:35:49 +0000 (14:35 +0900)]
Bug fix using same seed sequence for channel key geneartion
Matt Corallo [Wed, 31 Oct 2018 20:25:29 +0000 (16:25 -0400)]
Merge pull request #234 from TheBlueMatt/2018-10-169-redux
169 Tweaked
Schalk van Heerden [Wed, 31 Oct 2018 18:51:39 +0000 (14:51 -0400)]
Added config interface to allow users to specify channel limits
Matt Corallo [Wed, 31 Oct 2018 18:45:29 +0000 (14:45 -0400)]
Require at least 1k sat for outbound channels and meet BTC's dust
Matt Corallo [Wed, 31 Oct 2018 18:38:07 +0000 (14:38 -0400)]
Use non-funder's funding block wait instead of max with ours
This is both required by the protocol and also makes sense - if
we're the funder we don't mind accepting payment on the channel
after one confirmation because we assume we won't double-spend
ourselves.
Matt Corallo [Wed, 31 Oct 2018 19:05:19 +0000 (15:05 -0400)]
Merge pull request #232 from TheBlueMatt/master
Fix a full_stack_target crash
Matt Corallo [Mon, 29 Oct 2018 18:24:21 +0000 (14:24 -0400)]
[full_stack_target] don't allow double-confirming txn
Matt Corallo [Mon, 29 Oct 2018 17:38:15 +0000 (13:38 -0400)]
Give ManyChannelMonitor a logger and trace add_update events
Matt Corallo [Sat, 27 Oct 2018 14:49:05 +0000 (10:49 -0400)]
Merge pull request #226 from TheBlueMatt/2018-10-0.0.6
Bump version to 0.0.6
Matt Corallo [Sat, 27 Oct 2018 14:46:12 +0000 (10:46 -0400)]
Merge pull request #223 from TheBlueMatt/2018-10-chanmanager-serialize
Implement and document Channel/ChannelManager (de)serialization
Matt Corallo [Sat, 27 Oct 2018 14:46:04 +0000 (10:46 -0400)]
Merge pull request #222 from TheBlueMatt/2018-10-pre-commit-disconnect
Fix and test update_add_htlc but disconnect pre-commitment_signed
Matt Corallo [Sat, 27 Oct 2018 14:45:53 +0000 (10:45 -0400)]
Merge pull request #217 from TheBlueMatt/2018-10-msg-resp-overhaul
Send ChannelManager messages out-of-band to ensure ordered delivery
Matt Corallo [Sat, 27 Oct 2018 14:45:43 +0000 (10:45 -0400)]
Merge pull request #225 from TheBlueMatt/2018-10-214-redo
Keys Interface Simplification
Matt Corallo [Fri, 26 Oct 2018 21:24:46 +0000 (17:24 -0400)]
Bump version to 0.0.6
Matt Corallo [Fri, 26 Oct 2018 18:39:14 +0000 (14:39 -0400)]
Add very basic test of ChannelManager serialization round-trip
Matt Corallo [Sat, 27 Oct 2018 05:45:15 +0000 (01:45 -0400)]
fix broken state comparison (thanks @yuntai!)
Matt Corallo [Fri, 26 Oct 2018 18:35:50 +0000 (14:35 -0400)]
Implement and document Channel/ChannelManager (de)serialization
Matt Corallo [Fri, 26 Oct 2018 20:46:46 +0000 (16:46 -0400)]
Fix + test disconnect/reconnect prior to FundingLocked
Matt Corallo [Sat, 20 Oct 2018 22:48:14 +0000 (18:48 -0400)]
Fix and test update_add_htlc but disconnect pre-commitment_signed
Matt Corallo [Sat, 20 Oct 2018 22:17:19 +0000 (18:17 -0400)]
Avoid reentrancy of send_data from PeerHandler::read_bytes.
This greatly simplifies clients of PeerHandler, and because almost
all response messages have already been moved to process_events
this doesn't change much effeciency-wise.
Antoine Riard [Fri, 26 Oct 2018 17:41:07 +0000 (13:41 -0400)]
Integrate destination_script from KeysInterface in ChannelManager/Channel
Drop channel_monitor_claim_key from ChannelKeys
Matt Corallo [Thu, 25 Oct 2018 16:56:02 +0000 (12:56 -0400)]
Store+expose bits of Channel[Monitor] to figure out local state
Matt Corallo [Sat, 20 Oct 2018 21:50:34 +0000 (17:50 -0400)]
Send channel_reestablish out-of-band to ensure ordered deliver
Antoine Riard [Fri, 26 Oct 2018 17:35:57 +0000 (13:35 -0400)]
Integrate shutdown_pubkey from KeysInterface in ChannelManager/Channel
Drop channel_close_key from ChannelKeys
Matt Corallo [Wed, 24 Oct 2018 17:59:03 +0000 (13:59 -0400)]
Track ChannelMonitor-watched-outpoints (+ remove now-uesless Mutex)
Matt Corallo [Sat, 20 Oct 2018 21:18:53 +0000 (17:18 -0400)]
Respond to channel_reestablish out-of-band for ordered delivery
Antoine Riard [Fri, 26 Oct 2018 15:40:01 +0000 (11:40 -0400)]
Integrate KeysInterface to ChannelManager with Arc
Matt Corallo [Wed, 24 Oct 2018 15:14:12 +0000 (11:14 -0400)]
Track last_block_hash in ChannelMonitor and expose it on deser
Also make block_connected take a &mut self to ensure serialized
state will always be self-consistent.
Matt Corallo [Sat, 20 Oct 2018 16:56:42 +0000 (12:56 -0400)]
Send RAA/CS messages out-of-band to ensure ordered delivery
Antoine Riard [Fri, 26 Oct 2018 15:21:33 +0000 (11:21 -0400)]
Implement KeysInterface for KeysManager util
Matt Corallo [Wed, 24 Oct 2018 14:34:16 +0000 (10:34 -0400)]
Redo ChannelMonitor deserialization to avoid read_to_end()
This slightly changes the serialization format, but we're still
early enough that that's OK.
Matt Corallo [Sat, 20 Oct 2018 01:50:16 +0000 (21:50 -0400)]
Send shutdown/closing_signed msgs out-of-band for ordered delivery
Matt Corallo [Wed, 24 Oct 2018 14:33:42 +0000 (10:33 -0400)]
impl some additional (de)serializers, including a u48 wrapper type
Matt Corallo [Fri, 19 Oct 2018 21:30:52 +0000 (17:30 -0400)]
Send announcement_signatures msgs out-of-band for ordered delivery
Matt Corallo [Fri, 19 Oct 2018 01:34:41 +0000 (21:34 -0400)]
impl Readable/Writable for Route
Matt Corallo [Fri, 19 Oct 2018 21:06:40 +0000 (17:06 -0400)]
Send funding_signed messages out-of-band to ensure ordered delivery
Matt Corallo [Sat, 20 Oct 2018 22:46:03 +0000 (18:46 -0400)]
Add a BIG lock to ChannelManager
During normal operation we should never need to take this, so we
use a RwLock that allows normal parallelism until we want to
serialize out our ChannelManager, at which point we can take the
write-mode lock.
Matt Corallo [Fri, 19 Oct 2018 20:49:12 +0000 (16:49 -0400)]
Send AcceptChannel responses out-of-band to ensure ordered delivery
Matt Corallo [Thu, 18 Oct 2018 19:00:12 +0000 (15:00 -0400)]
Give ChannelMonitor a logger via new ReadableArgs trait
Matt Corallo [Fri, 19 Oct 2018 20:25:32 +0000 (16:25 -0400)]
Split Event, move MessageSendEvent push() inside channel_state lock
Matt Corallo [Thu, 18 Oct 2018 18:58:56 +0000 (14:58 -0400)]
Add a second Readable trait for state args, clean macros slightly
Matt Corallo [Thu, 18 Oct 2018 18:17:20 +0000 (14:17 -0400)]
Store [u8; 32]s instead of SharedSecrets (for deserialization)
Antoine Riard [Fri, 26 Oct 2018 15:15:55 +0000 (11:15 -0400)]
Add a KeysInterface which provides keys from user
Move ChannelKeys into keysinterface for generate a set of it from
master_seed and change return type to panic on bogus data
Antoine Riard [Fri, 19 Oct 2018 02:44:40 +0000 (02:44 +0000)]
Generate Events from ChannelMonitor to indicate spendable ouputs
Extend KeyStorage with delayed_payment_base_key and per_commitment_point
to derive local_delayed private key
Matt Corallo [Fri, 26 Oct 2018 01:21:54 +0000 (21:21 -0400)]
Merge pull request #224 from TheBlueMatt/2018-10-221-whitespace
#221 with a few trailing spaces removed
Yuntai Kyong [Thu, 25 Oct 2018 22:35:09 +0000 (07:35 +0900)]
Comments updated for PaymentReceived
Matt Corallo [Thu, 25 Oct 2018 19:19:06 +0000 (15:19 -0400)]
Merge pull request #220 from TheBlueMatt/2018-10-oops-xxx
Fix XXXs that slipped into router and handle HTLCFailCHannelUpdates
Matt Corallo [Thu, 25 Oct 2018 18:33:10 +0000 (14:33 -0400)]
Run travis-fuzz on Rust 1.29.2 due to 1.30 breaking honggfuzz build
Yuntai Kyong [Thu, 25 Oct 2018 08:21:41 +0000 (17:21 +0900)]
Add PaymentFailReason for fail_htlc_backwards
to indicate fail reason after a PaymentReceived event
Matt Corallo [Wed, 24 Oct 2018 04:44:15 +0000 (00:44 -0400)]
Fix XXXs that slipped into router and handle HTLCFailCHannelUpdates
Matt Corallo [Tue, 23 Oct 2018 20:45:50 +0000 (16:45 -0400)]
Merge pull request #213 from TheBlueMatt/2018-10-monitor-fail-pause
Add ChannelManager support for monitor update failure in one place
Matt Corallo [Thu, 18 Oct 2018 01:26:48 +0000 (21:26 -0400)]
Add basic sanity testing for channel monitor update failure handle
Matt Corallo [Wed, 17 Oct 2018 22:21:06 +0000 (18:21 -0400)]
ChannelManager support for monitor update failure in one place
Matt Corallo [Wed, 17 Oct 2018 22:19:55 +0000 (18:19 -0400)]
Add Channel support for monitor-update-failed pausing
Matt Corallo [Wed, 17 Oct 2018 15:35:26 +0000 (11:35 -0400)]
Add an Event for sending revoke_and_ack messages out-of-band
Matt Corallo [Thu, 18 Oct 2018 16:01:01 +0000 (12:01 -0400)]
Pull out the HTLC forwarding loop into a function
Matt Corallo [Wed, 17 Oct 2018 14:35:33 +0000 (10:35 -0400)]
Pull out regenerating RAA/CS in channel_reestablish handling
Matt Corallo [Wed, 17 Oct 2018 12:47:33 +0000 (08:47 -0400)]
Move channel monitor updates inside the channel_state lock
This really, really sucks as it defeats almost all of the
cross-channel parallelism we'd intended to have - waiting on a
client to update a watchtower for an unrelated channel to process
any messages is really shitty. We should revisit this with
per-channel locks as a compile-time option post-0.1.
Matt Corallo [Tue, 23 Oct 2018 20:03:08 +0000 (16:03 -0400)]
Merge pull request #212 from TheBlueMatt/2018-10-two-updates-disconnect
Fix reconnect message order on remote updates while waiting on RAA
Matt Corallo [Thu, 18 Oct 2018 01:17:01 +0000 (21:17 -0400)]
Add test for two-updates-in-flight disconnect channel resumption
Matt Corallo [Thu, 18 Oct 2018 02:13:31 +0000 (22:13 -0400)]
Fix reconnect message order on remote updates while waiting on RAA
Matt Corallo [Wed, 17 Oct 2018 22:06:13 +0000 (18:06 -0400)]
Add message ordering return value to handling channel_reestablish
Matt Corallo [Tue, 23 Oct 2018 19:25:07 +0000 (15:25 -0400)]
Merge pull request #219 from TheBlueMatt/2018-10-157-merge
Partially implement more onion error handling for sent payments
Yuntai Kyong [Mon, 22 Oct 2018 15:12:44 +0000 (11:12 -0400)]
Move HTLCFailChannelUpdate handling out-of-band
While this isn't neccessary for message ordering consistency, this
does mean that we won't end up processing an HTLCFailChannelUpdate
from a update_fail_htlc prior to it being fully committed (where if
the peer disconnects/reconnects it could theoretically give us a
different result, eg if their next-hop reconnected to them).
Yuntai Kyong [Sun, 14 Oct 2018 13:35:26 +0000 (22:35 +0900)]
Partially implement more onion error handling for sent payments
Also refactor out onion error handling into a function instead of
in update_fail_htlc.
Cache the initial htlc_msat that we sent instead of recalculating
it to check validity of the error returned.
Yuntai Kyong [Sun, 14 Oct 2018 13:34:45 +0000 (22:34 +0900)]
Some changes in comments and error messages.
Yuntai Kyong [Sun, 14 Oct 2018 13:30:21 +0000 (22:30 +0900)]
Error handling in decoding onion
Yuntai Kyong [Sun, 14 Oct 2018 13:11:21 +0000 (22:11 +0900)]
Add is_permanent field to ChannelClosed message and add NodeFailure
message
Matt Corallo [Tue, 16 Oct 2018 15:40:21 +0000 (11:40 -0400)]
Avoid needless on-chain channel failing for timing-out HTLCs
See new comments in code for more details
Matt Corallo [Thu, 18 Oct 2018 13:34:35 +0000 (09:34 -0400)]
Merge pull request #211 from TheBlueMatt/2018-10-misc-cleanups
Miscellaneous Cleanups
Matt Corallo [Wed, 17 Oct 2018 11:41:27 +0000 (07:41 -0400)]
Test double-reconnect
Matt Corallo [Mon, 15 Oct 2018 12:55:19 +0000 (08:55 -0400)]
Use APIError::ChannelUnavailable a bit more for consistency
Matt Corallo [Sun, 14 Oct 2018 20:37:21 +0000 (05:37 +0900)]
Ensure channel is live, not usable for sending fee updates
+ a freebie TODO removal
Matt Corallo [Sun, 14 Oct 2018 12:46:57 +0000 (21:46 +0900)]
Use live, not usable for list_usable_channels
Matt Corallo [Wed, 17 Oct 2018 01:05:17 +0000 (21:05 -0400)]
Merge pull request #210 from TheBlueMatt/2018-10-funding-disconnected
Fix + test funding tx confirmation while peer is disconnected
Matt Corallo [Tue, 16 Oct 2018 20:36:03 +0000 (16:36 -0400)]
Fix + test funding tx confirmation while peer is disconnected
Matt Corallo [Tue, 16 Oct 2018 17:17:28 +0000 (13:17 -0400)]
Merge pull request #208 from TheBlueMatt/2018-10-reconnect-fixes
Fix channel_reestablish exchanging in case of lost messages
Matt Corallo [Tue, 16 Oct 2018 03:04:59 +0000 (23:04 -0400)]
Add documentation about PaymentSent/Failed event duplication bug
Matt Corallo [Tue, 16 Oct 2018 02:11:52 +0000 (22:11 -0400)]
Test reconnecting after lost message(s) during the commitment dance
Matt Corallo [Tue, 16 Oct 2018 02:03:12 +0000 (22:03 -0400)]
Fix channel_reestablish exchanging in case of lost messages
This uses the new storage in HTLC state enums to reproduce the
various updates in a CommitmentUpdate group which is obviously
required to re-send a commitment_update after pending unreceived
updates were dropped.
Does not yet handle dropped update_fee updates properly.
Matt Corallo [Mon, 15 Oct 2018 19:11:02 +0000 (15:11 -0400)]
Store onion packet in OutboundHTLCState so we can regen update_add
Matt Corallo [Mon, 15 Oct 2018 18:38:19 +0000 (14:38 -0400)]
Store per-state HTLC data in the state enum itself (and store more)
Matt Corallo [Mon, 15 Oct 2018 17:03:44 +0000 (13:03 -0400)]
Split create_chan_between_nodes into two steps
Matt Corallo [Sat, 6 Oct 2018 03:49:26 +0000 (20:49 -0700)]
Merge pull request #206 from TheBlueMatt/2018-10-189-rebase
#189 rebased
Matt Corallo [Sat, 6 Oct 2018 02:55:55 +0000 (11:55 +0900)]
Further simplify get_{inbound,outbound}_pending_htlc_stats