Merge pull request #977 from TheBlueMatt/2021-06-fix-double-claim-close
[rust-lightning] / lightning / src / ln / channel.rs
index dc2dd758a9ea97bae465a69836f79200190c97d7..e08b8af49a3a40f5992f8e09b77f99f8279135fd 100644 (file)
@@ -301,6 +301,33 @@ pub struct CounterpartyForwardingInfo {
        pub cltv_expiry_delta: u16,
 }
 
+/// A return value enum for get_update_fulfill_htlc. See UpdateFulfillCommitFetch variants for
+/// description
+enum UpdateFulfillFetch {
+       NewClaim {
+               monitor_update: ChannelMonitorUpdate,
+               msg: Option<msgs::UpdateFulfillHTLC>,
+       },
+       DuplicateClaim {},
+}
+
+/// The return type of get_update_fulfill_htlc_and_commit.
+pub enum UpdateFulfillCommitFetch {
+       /// Indicates the HTLC fulfill is new, and either generated an update_fulfill message, placed
+       /// it in the holding cell, or re-generated the update_fulfill message after the same claim was
+       /// previously placed in the holding cell (and has since been removed).
+       NewClaim {
+               /// The ChannelMonitorUpdate which places the new payment preimage in the channel monitor
+               monitor_update: ChannelMonitorUpdate,
+               /// The update_fulfill message and commitment_signed message (if the claim was not placed
+               /// in the holding cell).
+               msgs: Option<(msgs::UpdateFulfillHTLC, msgs::CommitmentSigned)>,
+       },
+       /// Indicates the HTLC fulfill is duplicative and already existed either in the holding cell
+       /// or has been forgotten (presumably previously claimed).
+       DuplicateClaim {},
+}
+
 // TODO: We should refactor this to be an Inbound/OutboundChannel until initial setup handshaking
 // has been completed, and then turn into a Channel to get compiler-time enforcement of things like
 // calling channel_id() before we're set up or things like get_outbound_funding_signed on an
@@ -444,6 +471,15 @@ pub(super) struct Channel<Signer: Sign> {
        ///
        /// See-also <https://github.com/lightningnetwork/lnd/issues/4006>
        pub workaround_lnd_bug_4006: Option<msgs::FundingLocked>,
+
+       #[cfg(any(test, feature = "fuzztarget"))]
+       // When we receive an HTLC fulfill on an outbound path, we may immediately fulfill the
+       // corresponding HTLC on the inbound path. If, then, the outbound path channel is
+       // disconnected and reconnected (before we've exchange commitment_signed and revoke_and_ack
+       // messages), they may re-broadcast their update_fulfill_htlc, causing a duplicate claim. This
+       // is fine, but as a sanity check in our failure to generate the second claim, we check here
+       // that the original was a claim, and that we aren't now trying to fulfill a failed HTLC.
+       historical_inbound_htlc_fulfills: HashSet<u64>,
 }
 
 #[cfg(any(test, feature = "fuzztarget"))]
@@ -644,6 +680,9 @@ impl<Signer: Sign> Channel<Signer> {
                        next_remote_commitment_tx_fee_info_cached: Mutex::new(None),
 
                        workaround_lnd_bug_4006: None,
+
+                       #[cfg(any(test, feature = "fuzztarget"))]
+                       historical_inbound_htlc_fulfills: HashSet::new(),
                })
        }
 
@@ -889,6 +928,9 @@ impl<Signer: Sign> Channel<Signer> {
                        next_remote_commitment_tx_fee_info_cached: Mutex::new(None),
 
                        workaround_lnd_bug_4006: None,
+
+                       #[cfg(any(test, feature = "fuzztarget"))]
+                       historical_inbound_htlc_fulfills: HashSet::new(),
                };
 
                Ok(chan)
@@ -1216,13 +1258,7 @@ impl<Signer: Sign> Channel<Signer> {
                make_funding_redeemscript(&self.get_holder_pubkeys().funding_pubkey, self.counterparty_funding_pubkey())
        }
 
-       /// Per HTLC, only one get_update_fail_htlc or get_update_fulfill_htlc call may be made.
-       /// In such cases we debug_assert!(false) and return a ChannelError::Ignore. Thus, will always
-       /// return Ok(_) if debug assertions are turned on or preconditions are met.
-       ///
-       /// Note that it is still possible to hit these assertions in case we find a preimage on-chain
-       /// but then have a reorg which settles on an HTLC-failure on chain.
-       fn get_update_fulfill_htlc<L: Deref>(&mut self, htlc_id_arg: u64, payment_preimage_arg: PaymentPreimage, logger: &L) -> Result<(Option<msgs::UpdateFulfillHTLC>, Option<ChannelMonitorUpdate>), ChannelError> where L::Target: Logger {
+       fn get_update_fulfill_htlc<L: Deref>(&mut self, htlc_id_arg: u64, payment_preimage_arg: PaymentPreimage, logger: &L) -> UpdateFulfillFetch where L::Target: Logger {
                // Either ChannelFunded got set (which means it won't be unset) or there is no way any
                // caller thought we could have something claimed (cause we wouldn't have accepted in an
                // incoming HTLC anyway). If we got to ShutdownComplete, callers aren't allowed to call us,
@@ -1248,9 +1284,9 @@ impl<Signer: Sign> Channel<Signer> {
                                                if let &InboundHTLCRemovalReason::Fulfill(_) = reason {
                                                } else {
                                                        log_warn!(logger, "Have preimage and want to fulfill HTLC with payment hash {} we already failed against channel {}", log_bytes!(htlc.payment_hash.0), log_bytes!(self.channel_id()));
+                                                       debug_assert!(false, "Tried to fulfill an HTLC that was already failed");
                                                }
-                                               debug_assert!(false, "Tried to fulfill an HTLC that was already fail/fulfilled");
-                                               return Ok((None, None));
+                                               return UpdateFulfillFetch::DuplicateClaim {};
                                        },
                                        _ => {
                                                debug_assert!(false, "Have an inbound HTLC we tried to claim before it was fully committed to");
@@ -1262,7 +1298,11 @@ impl<Signer: Sign> Channel<Signer> {
                        }
                }
                if pending_idx == core::usize::MAX {
-                       return Err(ChannelError::Ignore("Unable to find a pending HTLC which matched the given HTLC ID".to_owned()));
+                       #[cfg(any(test, feature = "fuzztarget"))]
+                       // If we failed to find an HTLC to fulfill, make sure it was previously fulfilled and
+                       // this is simply a duplicate claim, not previously failed and we lost funds.
+                       debug_assert!(self.historical_inbound_htlc_fulfills.contains(&htlc_id_arg));
+                       return UpdateFulfillFetch::DuplicateClaim {};
                }
 
                // Now update local state:
@@ -1284,8 +1324,9 @@ impl<Signer: Sign> Channel<Signer> {
                                                if htlc_id_arg == htlc_id {
                                                        // Make sure we don't leave latest_monitor_update_id incremented here:
                                                        self.latest_monitor_update_id -= 1;
-                                                       debug_assert!(false, "Tried to fulfill an HTLC that was already fulfilled");
-                                                       return Ok((None, None));
+                                                       #[cfg(any(test, feature = "fuzztarget"))]
+                                                       debug_assert!(self.historical_inbound_htlc_fulfills.contains(&htlc_id_arg));
+                                                       return UpdateFulfillFetch::DuplicateClaim {};
                                                }
                                        },
                                        &HTLCUpdateAwaitingACK::FailHTLC { htlc_id, .. } => {
@@ -1294,7 +1335,7 @@ impl<Signer: Sign> Channel<Signer> {
                                                        // TODO: We may actually be able to switch to a fulfill here, though its
                                                        // rare enough it may not be worth the complexity burden.
                                                        debug_assert!(false, "Tried to fulfill an HTLC that was already failed");
-                                                       return Ok((None, Some(monitor_update)));
+                                                       return UpdateFulfillFetch::NewClaim { monitor_update, msg: None };
                                                }
                                        },
                                        _ => {}
@@ -1304,52 +1345,58 @@ impl<Signer: Sign> Channel<Signer> {
                        self.holding_cell_htlc_updates.push(HTLCUpdateAwaitingACK::ClaimHTLC {
                                payment_preimage: payment_preimage_arg, htlc_id: htlc_id_arg,
                        });
-                       return Ok((None, Some(monitor_update)));
+                       #[cfg(any(test, feature = "fuzztarget"))]
+                       self.historical_inbound_htlc_fulfills.insert(htlc_id_arg);
+                       return UpdateFulfillFetch::NewClaim { monitor_update, msg: None };
                }
+               #[cfg(any(test, feature = "fuzztarget"))]
+               self.historical_inbound_htlc_fulfills.insert(htlc_id_arg);
 
                {
                        let htlc = &mut self.pending_inbound_htlcs[pending_idx];
                        if let InboundHTLCState::Committed = htlc.state {
                        } else {
                                debug_assert!(false, "Have an inbound HTLC we tried to claim before it was fully committed to");
-                               return Ok((None, Some(monitor_update)));
+                               return UpdateFulfillFetch::NewClaim { monitor_update, msg: None };
                        }
                        log_trace!(logger, "Upgrading HTLC {} to LocalRemoved with a Fulfill in channel {}!", log_bytes!(htlc.payment_hash.0), log_bytes!(self.channel_id));
                        htlc.state = InboundHTLCState::LocalRemoved(InboundHTLCRemovalReason::Fulfill(payment_preimage_arg.clone()));
                }
 
-               Ok((Some(msgs::UpdateFulfillHTLC {
-                       channel_id: self.channel_id(),
-                       htlc_id: htlc_id_arg,
-                       payment_preimage: payment_preimage_arg,
-               }), Some(monitor_update)))
+               UpdateFulfillFetch::NewClaim {
+                       monitor_update,
+                       msg: Some(msgs::UpdateFulfillHTLC {
+                               channel_id: self.channel_id(),
+                               htlc_id: htlc_id_arg,
+                               payment_preimage: payment_preimage_arg,
+                       }),
+               }
        }
 
-       pub fn get_update_fulfill_htlc_and_commit<L: Deref>(&mut self, htlc_id: u64, payment_preimage: PaymentPreimage, logger: &L) -> Result<(Option<(msgs::UpdateFulfillHTLC, msgs::CommitmentSigned)>, Option<ChannelMonitorUpdate>), ChannelError> where L::Target: Logger {
-               match self.get_update_fulfill_htlc(htlc_id, payment_preimage, logger)? {
-                       (Some(update_fulfill_htlc), Some(mut monitor_update)) => {
-                               let (commitment, mut additional_update) = self.send_commitment_no_status_check(logger)?;
+       pub fn get_update_fulfill_htlc_and_commit<L: Deref>(&mut self, htlc_id: u64, payment_preimage: PaymentPreimage, logger: &L) -> Result<UpdateFulfillCommitFetch, (ChannelError, ChannelMonitorUpdate)> where L::Target: Logger {
+               match self.get_update_fulfill_htlc(htlc_id, payment_preimage, logger) {
+                       UpdateFulfillFetch::NewClaim { mut monitor_update, msg: Some(update_fulfill_htlc) } => {
+                               let (commitment, mut additional_update) = match self.send_commitment_no_status_check(logger) {
+                                       Err(e) => return Err((e, monitor_update)),
+                                       Ok(res) => res
+                               };
                                // send_commitment_no_status_check may bump latest_monitor_id but we want them to be
                                // strictly increasing by one, so decrement it here.
                                self.latest_monitor_update_id = monitor_update.update_id;
                                monitor_update.updates.append(&mut additional_update.updates);
-                               Ok((Some((update_fulfill_htlc, commitment)), Some(monitor_update)))
-                       },
-                       (Some(update_fulfill_htlc), None) => {
-                               let (commitment, monitor_update) = self.send_commitment_no_status_check(logger)?;
-                               Ok((Some((update_fulfill_htlc, commitment)), Some(monitor_update)))
+                               Ok(UpdateFulfillCommitFetch::NewClaim { monitor_update, msgs: Some((update_fulfill_htlc, commitment)) })
                        },
-                       (None, Some(monitor_update)) => Ok((None, Some(monitor_update))),
-                       (None, None) => Ok((None, None))
+                       UpdateFulfillFetch::NewClaim { monitor_update, msg: None } => Ok(UpdateFulfillCommitFetch::NewClaim { monitor_update, msgs: None }),
+                       UpdateFulfillFetch::DuplicateClaim {} => Ok(UpdateFulfillCommitFetch::DuplicateClaim {}),
                }
        }
 
-       /// Per HTLC, only one get_update_fail_htlc or get_update_fulfill_htlc call may be made.
-       /// In such cases we debug_assert!(false) and return a ChannelError::Ignore. Thus, will always
-       /// return Ok(_) if debug assertions are turned on or preconditions are met.
-       ///
-       /// Note that it is still possible to hit these assertions in case we find a preimage on-chain
-       /// but then have a reorg which settles on an HTLC-failure on chain.
+       /// We can only have one resolution per HTLC. In some cases around reconnect, we may fulfill
+       /// an HTLC more than once or fulfill once and then attempt to fail after reconnect. We cannot,
+       /// however, fail more than once as we wait for an upstream failure to be irrevocably committed
+       /// before we fail backwards.
+       /// If we do fail twice, we debug_assert!(false) and return Ok(None). Thus, will always return
+       /// Ok(_) if debug assertions are turned on or preconditions are met.
        pub fn get_update_fail_htlc<L: Deref>(&mut self, htlc_id_arg: u64, err_packet: msgs::OnionErrorPacket, logger: &L) -> Result<Option<msgs::UpdateFailHTLC>, ChannelError> where L::Target: Logger {
                if (self.channel_state & (ChannelState::ChannelFunded as u32)) != (ChannelState::ChannelFunded as u32) {
                        panic!("Was asked to fail an HTLC when channel was not in an operational state");
@@ -1365,8 +1412,11 @@ impl<Signer: Sign> Channel<Signer> {
                        if htlc.htlc_id == htlc_id_arg {
                                match htlc.state {
                                        InboundHTLCState::Committed => {},
-                                       InboundHTLCState::LocalRemoved(_) => {
-                                               debug_assert!(false, "Tried to fail an HTLC that was already fail/fulfilled");
+                                       InboundHTLCState::LocalRemoved(ref reason) => {
+                                               if let &InboundHTLCRemovalReason::Fulfill(_) = reason {
+                                               } else {
+                                                       debug_assert!(false, "Tried to fail an HTLC that was already failed");
+                                               }
                                                return Ok(None);
                                        },
                                        _ => {
@@ -1378,7 +1428,11 @@ impl<Signer: Sign> Channel<Signer> {
                        }
                }
                if pending_idx == core::usize::MAX {
-                       return Err(ChannelError::Ignore("Unable to find a pending HTLC which matched the given HTLC ID".to_owned()));
+                       #[cfg(any(test, feature = "fuzztarget"))]
+                       // If we failed to find an HTLC to fail, make sure it was previously fulfilled and this
+                       // is simply a duplicate fail, not previously failed and we failed-back too early.
+                       debug_assert!(self.historical_inbound_htlc_fulfills.contains(&htlc_id_arg));
+                       return Ok(None);
                }
 
                // Now update local state:
@@ -1387,8 +1441,9 @@ impl<Signer: Sign> Channel<Signer> {
                                match pending_update {
                                        &HTLCUpdateAwaitingACK::ClaimHTLC { htlc_id, .. } => {
                                                if htlc_id_arg == htlc_id {
-                                                       debug_assert!(false, "Tried to fail an HTLC that was already fulfilled");
-                                                       return Err(ChannelError::Ignore("Unable to find a pending HTLC which matched the given HTLC ID".to_owned()));
+                                                       #[cfg(any(test, feature = "fuzztarget"))]
+                                                       debug_assert!(self.historical_inbound_htlc_fulfills.contains(&htlc_id_arg));
+                                                       return Ok(None);
                                                }
                                        },
                                        &HTLCUpdateAwaitingACK::FailHTLC { htlc_id, .. } => {
@@ -2435,24 +2490,28 @@ impl<Signer: Sign> Channel<Signer> {
                                                }
                                        },
                                        &HTLCUpdateAwaitingACK::ClaimHTLC { ref payment_preimage, htlc_id, .. } => {
-                                               match self.get_update_fulfill_htlc(htlc_id, *payment_preimage, logger) {
-                                                       Ok((update_fulfill_msg_option, additional_monitor_update_opt)) => {
-                                                               update_fulfill_htlcs.push(update_fulfill_msg_option.unwrap());
-                                                               if let Some(mut additional_monitor_update) = additional_monitor_update_opt {
-                                                                       monitor_update.updates.append(&mut additional_monitor_update.updates);
-                                                               }
-                                                       },
-                                                       Err(e) => {
-                                                               if let ChannelError::Ignore(_) = e {}
-                                                               else {
-                                                                       panic!("Got a non-IgnoreError action trying to fulfill holding cell HTLC");
-                                                               }
-                                                       }
-                                               }
+                                               // If an HTLC claim was previously added to the holding cell (via
+                                               // `get_update_fulfill_htlc`, then generating the claim message itself must
+                                               // not fail - any in between attempts to claim the HTLC will have resulted
+                                               // in it hitting the holding cell again and we cannot change the state of a
+                                               // holding cell HTLC from fulfill to anything else.
+                                               let (update_fulfill_msg_option, mut additional_monitor_update) =
+                                                       if let UpdateFulfillFetch::NewClaim { msg, monitor_update } = self.get_update_fulfill_htlc(htlc_id, *payment_preimage, logger) {
+                                                               (msg, monitor_update)
+                                                       } else { unreachable!() };
+                                               update_fulfill_htlcs.push(update_fulfill_msg_option.unwrap());
+                                               monitor_update.updates.append(&mut additional_monitor_update.updates);
                                        },
                                        &HTLCUpdateAwaitingACK::FailHTLC { htlc_id, ref err_packet } => {
                                                match self.get_update_fail_htlc(htlc_id, err_packet.clone(), logger) {
-                                                       Ok(update_fail_msg_option) => update_fail_htlcs.push(update_fail_msg_option.unwrap()),
+                                                       Ok(update_fail_msg_option) => {
+                                                               // If an HTLC failure was previously added to the holding cell (via
+                                                               // `get_update_fail_htlc`) then generating the fail message itself
+                                                               // must not fail - we should never end up in a state where we
+                                                               // double-fail an HTLC or fail-then-claim an HTLC as it indicates
+                                                               // we didn't wait for a full revocation before failing.
+                                                               update_fail_htlcs.push(update_fail_msg_option.unwrap())
+                                                       },
                                                        Err(e) => {
                                                                if let ChannelError::Ignore(_) = e {}
                                                                else {
@@ -4684,6 +4743,13 @@ impl<Signer: Sign> Writeable for Channel<Signer> {
 
                self.channel_update_status.write(writer)?;
 
+               #[cfg(any(test, feature = "fuzztarget"))]
+               (self.historical_inbound_htlc_fulfills.len() as u64).write(writer)?;
+               #[cfg(any(test, feature = "fuzztarget"))]
+               for htlc in self.historical_inbound_htlc_fulfills.iter() {
+                       htlc.write(writer)?;
+               }
+
                write_tlv_fields!(writer, {
                        (0, self.announcement_sigs, option),
                        // minimum_depth and counterparty_selected_channel_reserve_satoshis used to have a
@@ -4893,6 +4959,16 @@ impl<'a, Signer: Sign, K: Deref> ReadableArgs<&'a K> for Channel<Signer>
 
                let channel_update_status = Readable::read(reader)?;
 
+               #[cfg(any(test, feature = "fuzztarget"))]
+               let mut historical_inbound_htlc_fulfills = HashSet::new();
+               #[cfg(any(test, feature = "fuzztarget"))]
+               {
+                       let htlc_fulfills_len: u64 = Readable::read(reader)?;
+                       for _ in 0..htlc_fulfills_len {
+                               assert!(historical_inbound_htlc_fulfills.insert(Readable::read(reader)?));
+                       }
+               }
+
                let mut announcement_sigs = None;
                read_tlv_fields!(reader, {
                        (0, announcement_sigs, option),
@@ -4985,6 +5061,9 @@ impl<'a, Signer: Sign, K: Deref> ReadableArgs<&'a K> for Channel<Signer>
                        next_remote_commitment_tx_fee_info_cached: Mutex::new(None),
 
                        workaround_lnd_bug_4006: None,
+
+                       #[cfg(any(test, feature = "fuzztarget"))]
+                       historical_inbound_htlc_fulfills,
                })
        }
 }