]> git.bitcoin.ninja Git - rust-lightning/commitdiff
Avoid generating redundant claims after initial confirmation
authorWilmer Paulino <wilmer.paulino@gmail.com>
Wed, 21 Sep 2022 19:54:28 +0000 (12:54 -0700)
committerWilmer Paulino <wilmer.paulino@gmail.com>
Wed, 2 Nov 2022 17:07:45 +0000 (10:07 -0700)
These claims will never be valid as a previous claim has already
confirmed. If a previous claim is reorged out of the chain, a new claim
will be generated bypassing the new behavior.

While this doesn't change much for our existing transaction-based
claims, as broadcasting an already confirmed transaction acts as a NOP,
it prevents us from yielding redundant event-based claims, which will be
introduced as part of the anchors patchset.

lightning/src/chain/onchaintx.rs
lightning/src/ln/functional_tests.rs

index 18679f0aacd5cd34fead7952fc91be997ad4e57b..9cfa9d288ce3590a585a3e7e708e0bcb54911b30 100644 (file)
@@ -430,7 +430,43 @@ impl<ChannelSigner: Sign> OnchainTxHandler<ChannelSigner> {
                where F::Target: FeeEstimator,
                                        L::Target: Logger,
        {
-               if cached_request.outpoints().len() == 0 { return None } // But don't prune pending claiming request yet, we may have to resurrect HTLCs
+               let request_outpoints = cached_request.outpoints();
+               if request_outpoints.is_empty() {
+                       // Don't prune pending claiming request yet, we may have to resurrect HTLCs. Untractable
+                       // packages cannot be aggregated and will never be split, so we cannot end up with an
+                       // empty claim.
+                       debug_assert!(cached_request.is_malleable());
+                       return None;
+               }
+               // If we've seen transaction inclusion in the chain for all outpoints in our request, we
+               // don't need to continue generating more claims. We'll keep tracking the request to fully
+               // remove it once it reaches the confirmation threshold, or to generate a new claim if the
+               // transaction is reorged out.
+               let mut all_inputs_have_confirmed_spend = true;
+               for outpoint in &request_outpoints {
+                       if let Some(first_claim_txid_height) = self.claimable_outpoints.get(outpoint) {
+                               // We check for outpoint spends within claims individually rather than as a set
+                               // since requests can have outpoints split off.
+                               if !self.onchain_events_awaiting_threshold_conf.iter()
+                                       .any(|event_entry| if let OnchainEvent::Claim { claim_request } = event_entry.event {
+                                               first_claim_txid_height.0 == claim_request
+                                       } else {
+                                               // The onchain event is not a claim, keep seeking until we find one.
+                                               false
+                                       })
+                               {
+                                       // Either we had no `OnchainEvent::Claim`, or we did but none matched the
+                                       // outpoint's registered spend.
+                                       all_inputs_have_confirmed_spend = false;
+                               }
+                       } else {
+                               // The request's outpoint spend does not exist yet.
+                               all_inputs_have_confirmed_spend = false;
+                       }
+               }
+               if all_inputs_have_confirmed_spend {
+                       return None;
+               }
 
                // Compute new height timer to decide when we need to regenerate a new bumped version of the claim tx (if we
                // didn't receive confirmation of it before, or not enough reorg-safe depth on top of it).
index eb6d7032fee67d7714005f807d2ac121c48a9a8d..a2e59b58c871c7212282122ae6162046c081b342 100644 (file)
@@ -2951,26 +2951,8 @@ fn do_test_htlc_on_chain_timeout(connect_style: ConnectStyle) {
        mine_transaction(&nodes[1], &timeout_tx);
        check_added_monitors!(nodes[1], 1);
        check_closed_broadcast!(nodes[1], true);
-       {
-               // B will rebroadcast a fee-bumped timeout transaction here.
-               let node_txn = nodes[1].tx_broadcaster.txn_broadcasted.lock().unwrap().split_off(0);
-               assert_eq!(node_txn.len(), 1);
-               check_spends!(node_txn[0], commitment_tx[0]);
-       }
 
        connect_blocks(&nodes[1], ANTI_REORG_DELAY - 1);
-       {
-               // B may rebroadcast its own holder commitment transaction here, as a safeguard against
-               // some incredibly unlikely partial-eclipse-attack scenarios. That said, because the
-               // original commitment_tx[0] (also spending chan_2.3) has reached ANTI_REORG_DELAY B really
-               // shouldn't broadcast anything here, and in some connect style scenarios we do not.
-               let node_txn = nodes[1].tx_broadcaster.txn_broadcasted.lock().unwrap().split_off(0);
-               if node_txn.len() == 1 {
-                       check_spends!(node_txn[0], chan_2.3);
-               } else {
-                       assert_eq!(node_txn.len(), 0);
-               }
-       }
 
        expect_pending_htlcs_forwardable_and_htlc_handling_failed!(nodes[1], vec![HTLCDestination::NextHopChannel { node_id: Some(nodes[2].node.get_our_node_id()), channel_id: chan_2.2 }]);
        check_added_monitors!(nodes[1], 1);
@@ -8001,22 +7983,6 @@ fn test_bump_penalty_txn_on_revoked_htlcs() {
        connect_block(&nodes[0], &Block { header: header_130, txdata: penalty_txn });
        let header_131 = BlockHeader { version: 0x20000000, prev_blockhash: header_130.block_hash(), merkle_root: TxMerkleNode::all_zeros(), time: 42, bits: 42, nonce: 42 };
        connect_block(&nodes[0], &Block { header: header_131, txdata: Vec::new() });
-       {
-               let mut node_txn = nodes[0].tx_broadcaster.txn_broadcasted.lock().unwrap();
-               assert_eq!(node_txn.len(), 2); // 2 bumped penalty txn on revoked commitment tx
-
-               check_spends!(node_txn[0], revoked_local_txn[0]);
-               check_spends!(node_txn[1], revoked_local_txn[0]);
-               // Note that these are both bogus - they spend outputs already claimed in block 129:
-               if node_txn[0].input[0].previous_output == revoked_htlc_txn[0].input[0].previous_output  {
-                       assert_eq!(node_txn[1].input[0].previous_output, revoked_htlc_txn[2].input[0].previous_output);
-               } else {
-                       assert_eq!(node_txn[0].input[0].previous_output, revoked_htlc_txn[2].input[0].previous_output);
-                       assert_eq!(node_txn[1].input[0].previous_output, revoked_htlc_txn[0].input[0].previous_output);
-               }
-
-               node_txn.clear();
-       };
 
        // Few more blocks to confirm penalty txn
        connect_blocks(&nodes[0], 4);