Merge #18044: Use wtxid for transaction relay

0a4f1422cd1c20e12a05d7ff1a2ef1d5e7c654bb Further improve comments around recentRejects (Suhas Daftuar)
0e20cfedb704c1f76bb727e2009867d3d503a03d Disconnect peers sending wtxidrelay message after VERACK (Suhas Daftuar)
cacd85209e20fc0fd08f86eed23b6ef93484ffcf test: Use wtxid relay generally in functional tests (Fabian Jahr)
8d8099e97ab8af2126f6fbd223fbd82c52f2e85e test: Add tests for wtxid tx relay in segwit test (Fabian Jahr)
9a5392fdf67f1c5c90f52d3cdb3dea4f35d1609f test: Update test framework p2p protocol version to 70016 (Fabian Jahr)
dd78d1d641178b473ab1156b71a837b9e686792b Rename AddInventoryKnown() to AddKnownTx() (Suhas Daftuar)
4eb515574e1012bc8ea5dafc3042dcdf4c766f26 Make TX_WITNESS_STRIPPED its own rejection reason (Suhas Daftuar)
97141ca442daea8fc9c307cf81a02b38dcc28fd8 Delay getdata requests from peers using txid-based relay (Suhas Daftuar)
46d78d47dea345329ba094310eec56ab00a02ddc Add p2p message "wtxidrelay" (Suhas Daftuar)
2d282e0cba9761574b6b43d134ca95f3052d7fd2 ignore non-wtxidrelay compliant invs (Anthony Towns)
ac88e2eb619821ad7ae1d45d4b40be69051d3999 Add support for tx-relay via wtxid (Suhas Daftuar)
8e68fc246d09f1e6c6dfa8c676969d97c2eb4334 Add wtxids to recentRejects instead of txids (Suhas Daftuar)
144c38582042c3b9ec53bb97ba0644fc0114f8fb Add wtxids of confirmed transactions to bloom filter (Suhas Daftuar)
85c78d54af462996a0bca6cf97f91e1aa8778ae8 Add wtxid-index to orphan map (Suhas Daftuar)
08b39955ec7f84e835ab0b1366f0dd28dfd6ce03 Add a wtxid-index to mapRelay (Suhas Daftuar)
60f0acda713e7b9dc188aef54ef93981a93f4e44 Just pass a hash to AddInventoryKnown (Suhas Daftuar)
c7eb6b4f1fe5bd76388a023529977674534334a7 Add wtxid to mempool unbroadcast tracking (Amiti Uttarwar)
2b4b90aa8f0440deacefb5997d7bd1f9f5c591b3 Add a wtxid-index to the mempool (Suhas Daftuar)

Pull request description:

  Using txids (a transaction's hash, without witness) for transaction relay is problematic, post-segwit -- if a peer gives us a segwit transaction that fails policy checks, it could be because the txid associated with the transaction is definitely unacceptable to our node (regardless of the witness), or it could be that the transaction was malleated and with a different witness, the txid could be accepted to our mempool.

  We have a bloom filter of recently rejected transactions, whose purpose is to help us avoid redownloading and revalidating transactions that fail to be accepted, but because of this potential for witness malleability to interfere with relay of valid transactions, we do not use the filter for segwit transactions.  This issue is discussed at some length in #8279.  The effect of this is that whenever a segwit transaction that fails policy checks is relayed, a node would download that transaction from every peer announcing it, because it has no way presently to cache failure.  Historically this hasn't been a big problem, but if/when policy for accepting segwit transactions were to change (eg taproot, or any other change), we could expect older nodes talking to newer nodes to be wasting bandwidth because of this.

  As discussed in that issue, switching to wtxid-based relay solves this problem -- by using an identifier for a transaction that commits to all the data in our relay protocol, we can be certain if a transaction that a peer is announcing is one that we've already tried to process, or if it's something new.  This PR introduces support for wtxid-based relay with peers that support it (and remains backwards compatible with peers that use txids for relay, of course).

  Apart from code correctness, one issue to be aware of is that by downloading from old and new peers alike, we should expect there to be some bandwidth wasted, because sometimes we might download the same transaction via txid-relay as well as wtxid-relay.  The last commit in this PR implements a heuristic I want to analyze, which is to just delay relay from txid-relay peers by 2 seconds, if we have at least 1 wtxid-based peer.  I've just started running a couple nodes with this heuristic so I can measure how well it works, but I'm open to other ideas for minimizing that issue.  In the long run, I think this will be essentially a non-issue, so I don't think it's too big a concern, we just need to bite the bullet and deal with it during upgrade.

  Finally, this proposal would need a simple BIP describing the changes, which I haven't yet drafted.  However, review and testing of this code in the interim would be welcome.

  To do items:
  - [x] Write BIP explaining the spec here (1 new p2p message for negotiating wtxid-based relay, along with a new INV type)
  - [ ] Measure and evaluate a heuristic for minimizing how often a node downloads the same transaction twice, when connected to old and new nodes.

ACKs for top commit:
  naumenkogs:
    utACK 0a4f1422cd1c20e12a05d7ff1a2ef1d5e7c654bb
  laanwj:
    utACK 0a4f1422cd1c20e12a05d7ff1a2ef1d5e7c654bb

Tree-SHA512: d8eb8f0688cf0cbe9507bf738e143edab1f595551fdfeddc2b6734686ea26e7f156b6bfde38bad8bbbe8bec1857c7223e1687f8f018de7463dde8ecaa8f450df

18 files changed, 450 insertions, 114 deletions

diff --git a/src/consensus/validation.h b/src/consensus/validation.h
index 74d6138041..8de7a8f2d8 100644
--- a/src/consensus/validation.h
+++ b/src/consensus/validation.h
@@ -30,12 +30,16 @@ enum class TxValidationResult {
     TX_MISSING_INPUTS,        //!< transaction was missing some of its inputs
     TX_PREMATURE_SPEND,       //!< transaction spends a coinbase too early, or violates locktime/sequence locks
     /**
-     * Transaction might be missing a witness, have a witness prior to SegWit
+     * Transaction might have a witness prior to SegWit
      * activation, or witness may have been malleated (which includes
      * non-standard witnesses).
      */
     TX_WITNESS_MUTATED,
     /**
+     * Transaction is missing a witness.
+     */
+    TX_WITNESS_STRIPPED,
+    /**
      * Tx already in mempool or conflicts with a tx in the chain
      * (if it conflicts with another tx in mempool, we use MEMPOOL_POLICY as it failed to reach the RBF threshold)
      * Currently this is only used if the transaction already exists in the mempool or on chain.
diff --git a/src/net.h b/src/net.h
index d980986c44..72c73dc3a1 100644
--- a/src/net.h
+++ b/src/net.h
@@ -965,11 +965,11 @@ public:
     }
 
 
-    void AddInventoryKnown(const CInv& inv)
+    void AddKnownTx(const uint256& hash)
     {
         if (m_tx_relay != nullptr) {
             LOCK(m_tx_relay->cs_tx_inventory);
-            m_tx_relay->filterInventoryKnown.insert(inv.hash);
+            m_tx_relay->filterInventoryKnown.insert(hash);
         }
     }
 
diff --git a/src/net_processing.cpp b/src/net_processing.cpp
index 7a58de35d7..5f1e7318f3 100644
--- a/src/net_processing.cpp
+++ b/src/net_processing.cpp
@@ -75,6 +75,8 @@ static const unsigned int MAX_INV_SZ = 50000;
 static constexpr int32_t MAX_PEER_TX_IN_FLIGHT = 100;
 /** Maximum number of announced transactions from a peer */
 static constexpr int32_t MAX_PEER_TX_ANNOUNCEMENTS = 2 * MAX_INV_SZ;
+/** How many microseconds to delay requesting transactions via txids, if we have wtxid-relaying peers */
+static constexpr std::chrono::microseconds TXID_RELAY_DELAY{std::chrono::seconds{2}};
 /** How many microseconds to delay requesting transactions from inbound peers */
 static constexpr std::chrono::microseconds INBOUND_PEER_TX_DELAY{std::chrono::seconds{2}};
 /** How long to wait (in microseconds) before downloading a transaction from an additional peer */
@@ -151,6 +153,7 @@ struct COrphanTx {
 };
 RecursiveMutex g_cs_orphans;
 std::map<uint256, COrphanTx> mapOrphanTransactions GUARDED_BY(g_cs_orphans);
+std::map<uint256, std::map<uint256, COrphanTx>::iterator> g_orphans_by_wtxid GUARDED_BY(g_cs_orphans);
 
 void EraseOrphansFor(NodeId peer);
 
@@ -187,6 +190,15 @@ namespace {
      * million to make it highly unlikely for users to have issues with this
      * filter.
      *
+     * We only need to add wtxids to this filter. For non-segwit
+     * transactions, the txid == wtxid, so this only prevents us from
+     * re-downloading non-segwit transactions when communicating with
+     * non-wtxidrelay peers -- which is important for avoiding malleation
+     * attacks that could otherwise interfere with transaction relay from
+     * non-wtxidrelay peers. For communicating with wtxidrelay peers, having
+     * the reject filter store wtxids is exactly what we want to avoid
+     * redownload of a rejected transaction.
+     *
      * Memory used: 1.3 MB
      */
     std::unique_ptr<CRollingBloomFilter> recentRejects GUARDED_BY(cs_main);
@@ -218,6 +230,9 @@ namespace {
     /** Number of peers from which we're downloading blocks. */
     int nPeersWithValidatedDownloads GUARDED_BY(cs_main) = 0;
 
+    /** Number of peers with wtxid relay. */
+    int g_wtxid_relay_peers GUARDED_BY(cs_main) = 0;
+
     /** Number of outbound peers with m_chain_sync.m_protect. */
     int g_outbound_peers_with_protect_from_disconnect GUARDED_BY(cs_main) = 0;
 
@@ -409,6 +424,9 @@ struct CNodeState {
     //! A rolling bloom filter of all announced tx CInvs to this peer.
     CRollingBloomFilter m_recently_announced_invs = CRollingBloomFilter{INVENTORY_MAX_RECENT_RELAY, 0.000001};
 
+    //! Whether this peer relays txs via wtxid
+    bool m_wtxid_relay{false};
+
     CNodeState(CAddress addrIn, std::string addrNameIn, bool is_inbound, bool is_manual) :
         address(addrIn), name(std::move(addrNameIn)), m_is_inbound(is_inbound),
         m_is_manual_connection (is_manual)
@@ -760,7 +778,7 @@ void UpdateTxRequestTime(const uint256& txid, std::chrono::microseconds request_
     }
 }
 
-std::chrono::microseconds CalculateTxGetDataTime(const uint256& txid, std::chrono::microseconds current_time, bool use_inbound_delay) EXCLUSIVE_LOCKS_REQUIRED(cs_main)
+std::chrono::microseconds CalculateTxGetDataTime(const uint256& txid, std::chrono::microseconds current_time, bool use_inbound_delay, bool use_txid_delay) EXCLUSIVE_LOCKS_REQUIRED(cs_main)
 {
     std::chrono::microseconds process_time;
     const auto last_request_time = GetTxRequestTime(txid);
@@ -776,6 +794,9 @@ std::chrono::microseconds CalculateTxGetDataTime(const uint256& txid, std::chron
     // We delay processing announcements from inbound peers
     if (use_inbound_delay) process_time += INBOUND_PEER_TX_DELAY;
 
+    // We delay processing announcements from peers that use txid-relay (instead of wtxid)
+    if (use_txid_delay) process_time += TXID_RELAY_DELAY;
+
     return process_time;
 }
 
@@ -793,7 +814,7 @@ void RequestTx(CNodeState* state, const uint256& txid, std::chrono::microseconds
 
     // Calculate the time to try requesting this transaction. Use
     // fPreferredDownload as a proxy for outbound peers.
-    const auto process_time = CalculateTxGetDataTime(txid, current_time, !state->fPreferredDownload);
+    const auto process_time = CalculateTxGetDataTime(txid, current_time, !state->fPreferredDownload, !state->m_wtxid_relay && g_wtxid_relay_peers > 0);
 
     peer_download_state.m_tx_process_time.emplace(process_time, txid);
 }
@@ -830,14 +851,15 @@ void PeerLogicValidation::InitializeNode(CNode *pnode) {
 
 void PeerLogicValidation::ReattemptInitialBroadcast(CScheduler& scheduler) const
 {
-    std::set<uint256> unbroadcast_txids = m_mempool.GetUnbroadcastTxs();
+    std::map<uint256, uint256> unbroadcast_txids = m_mempool.GetUnbroadcastTxs();
 
-    for (const uint256& txid : unbroadcast_txids) {
+    for (const auto& elem : unbroadcast_txids) {
         // Sanity check: all unbroadcast txns should exist in the mempool
-        if (m_mempool.exists(txid)) {
-            RelayTransaction(txid, *connman);
+        if (m_mempool.exists(elem.first)) {
+            LOCK(cs_main);
+            RelayTransaction(elem.first, elem.second, *connman);
         } else {
-            m_mempool.RemoveUnbroadcastTx(txid, true);
+            m_mempool.RemoveUnbroadcastTx(elem.first, true);
         }
     }
 
@@ -869,6 +891,8 @@ void PeerLogicValidation::FinalizeNode(NodeId nodeid, bool& fUpdateConnectionTim
     assert(nPeersWithValidatedDownloads >= 0);
     g_outbound_peers_with_protect_from_disconnect -= state->m_chain_sync.m_protect;
     assert(g_outbound_peers_with_protect_from_disconnect >= 0);
+    g_wtxid_relay_peers -= state->m_wtxid_relay;
+    assert(g_wtxid_relay_peers >= 0);
 
     mapNodeState.erase(nodeid);
 
@@ -878,6 +902,7 @@ void PeerLogicValidation::FinalizeNode(NodeId nodeid, bool& fUpdateConnectionTim
         assert(nPreferredDownload == 0);
         assert(nPeersWithValidatedDownloads == 0);
         assert(g_outbound_peers_with_protect_from_disconnect == 0);
+        assert(g_wtxid_relay_peers == 0);
     }
     LogPrint(BCLog::NET, "Cleared nodestate for peer=%d\n", nodeid);
 }
@@ -936,6 +961,8 @@ bool AddOrphanTx(const CTransactionRef& tx, NodeId peer) EXCLUSIVE_LOCKS_REQUIRE
     auto ret = mapOrphanTransactions.emplace(hash, COrphanTx{tx, peer, GetTime() + ORPHAN_TX_EXPIRE_TIME, g_orphan_list.size()});
     assert(ret.second);
     g_orphan_list.push_back(ret.first);
+    // Allow for lookups in the orphan pool by wtxid, as well as txid
+    g_orphans_by_wtxid.emplace(tx->GetWitnessHash(), ret.first);
     for (const CTxIn& txin : tx->vin) {
         mapOrphanTransactionsByPrev[txin.prevout].insert(ret.first);
     }
@@ -972,6 +999,7 @@ int static EraseOrphanTx(uint256 hash) EXCLUSIVE_LOCKS_REQUIRED(g_cs_orphans)
         it_last->second.list_pos = old_pos;
     }
     g_orphan_list.pop_back();
+    g_orphans_by_wtxid.erase(it->second.tx->GetWitnessHash());
 
     mapOrphanTransactions.erase(it);
     return 1;
@@ -1141,6 +1169,7 @@ static bool MaybePunishNodeForTx(NodeId nodeid, const TxValidationState& state,
     case TxValidationResult::TX_MISSING_INPUTS:
     case TxValidationResult::TX_PREMATURE_SPEND:
     case TxValidationResult::TX_WITNESS_MUTATED:
+    case TxValidationResult::TX_WITNESS_STRIPPED:
     case TxValidationResult::TX_CONFLICT:
     case TxValidationResult::TX_MEMPOOL_POLICY:
         break;
@@ -1181,14 +1210,15 @@ PeerLogicValidation::PeerLogicValidation(CConnman* connmanIn, BanMan* banman, CS
     recentRejects.reset(new CRollingBloomFilter(120000, 0.000001));
 
     // Blocks don't typically have more than 4000 transactions, so this should
-    // be at least six blocks (~1 hr) worth of transactions that we can store.
+    // be at least six blocks (~1 hr) worth of transactions that we can store,
+    // inserting both a txid and wtxid for every observed transaction.
     // If the number of transactions appearing in a block goes up, or if we are
     // seeing getdata requests more than an hour after initial announcement, we
     // can increase this number.
     // The false positive rate of 1/1M should come out to less than 1
     // transaction per day that would be inadvertently ignored (which is the
     // same probability that we have in the reject filter).
-    g_recent_confirmed_transactions.reset(new CRollingBloomFilter(24000, 0.000001));
+    g_recent_confirmed_transactions.reset(new CRollingBloomFilter(48000, 0.000001));
 
     const Consensus::Params& consensusParams = Params().GetConsensus();
     // Stale tip checking and peer eviction are on two different timers, but we
@@ -1244,6 +1274,9 @@ void PeerLogicValidation::BlockConnected(const std::shared_ptr<const CBlock>& pb
         LOCK(g_cs_recent_confirmed_transactions);
         for (const auto& ptx : pblock->vtx) {
             g_recent_confirmed_transactions->insert(ptx->GetHash());
+            if (ptx->GetHash() != ptx->GetWitnessHash()) {
+                g_recent_confirmed_transactions->insert(ptx->GetWitnessHash());
+            }
         }
     }
 }
@@ -1397,6 +1430,7 @@ bool static AlreadyHave(const CInv& inv, const CTxMemPool& mempool) EXCLUSIVE_LO
     {
     case MSG_TX:
     case MSG_WITNESS_TX:
+    case MSG_WTX:
         {
             assert(recentRejects);
             if (::ChainActive().Tip()->GetBlockHash() != hashRecentRejectsChainTip)
@@ -1411,7 +1445,11 @@ bool static AlreadyHave(const CInv& inv, const CTxMemPool& mempool) EXCLUSIVE_LO
 
             {
                 LOCK(g_cs_orphans);
-                if (mapOrphanTransactions.count(inv.hash)) return true;
+                if (inv.type != MSG_WTX && mapOrphanTransactions.count(inv.hash)) {
+                    return true;
+                } else if (inv.type == MSG_WTX && g_orphans_by_wtxid.count(inv.hash)) {
+                    return true;
+                }
             }
 
             {
@@ -1419,8 +1457,8 @@ bool static AlreadyHave(const CInv& inv, const CTxMemPool& mempool) EXCLUSIVE_LO
                 if (g_recent_confirmed_transactions->contains(inv.hash)) return true;
             }
 
-            return recentRejects->contains(inv.hash) ||
-                   mempool.exists(inv.hash);
+            const bool by_wtxid = (inv.type == MSG_WTX);
+            return recentRejects->contains(inv.hash) || mempool.exists(inv.hash, by_wtxid);
         }
     case MSG_BLOCK:
     case MSG_WITNESS_BLOCK:
@@ -1430,11 +1468,17 @@ bool static AlreadyHave(const CInv& inv, const CTxMemPool& mempool) EXCLUSIVE_LO
     return true;
 }
 
-void RelayTransaction(const uint256& txid, const CConnman& connman)
+void RelayTransaction(const uint256& txid, const uint256& wtxid, const CConnman& connman)
 {
-    connman.ForEachNode([&txid](CNode* pnode)
+    connman.ForEachNode([&txid, &wtxid](CNode* pnode)
     {
-        pnode->PushTxInventory(txid);
+        AssertLockHeld(cs_main);
+        CNodeState &state = *State(pnode->GetId());
+        if (state.m_wtxid_relay) {
+            pnode->PushTxInventory(wtxid);
+        } else {
+            pnode->PushTxInventory(txid);
+        }
     });
 }
 
@@ -1632,9 +1676,9 @@ void static ProcessGetBlockData(CNode& pfrom, const CChainParams& chainparams, c
 }
 
 //! Determine whether or not a peer can request a transaction, and return it (or nullptr if not found or not allowed).
-CTransactionRef static FindTxForGetData(const CNode& peer, const uint256& txid, const std::chrono::seconds mempool_req, const std::chrono::seconds now) LOCKS_EXCLUDED(cs_main)
+CTransactionRef static FindTxForGetData(const CNode& peer, const uint256& txid_or_wtxid, bool use_wtxid, const std::chrono::seconds mempool_req, const std::chrono::seconds now) LOCKS_EXCLUDED(cs_main)
 {
-    auto txinfo = mempool.info(txid);
+    auto txinfo = mempool.info(txid_or_wtxid, use_wtxid);
     if (txinfo.tx) {
         // If a TX could have been INVed in reply to a MEMPOOL request,
         // or is older than UNCONDITIONAL_RELAY_DELAY, permit the request
@@ -1646,13 +1690,12 @@ CTransactionRef static FindTxForGetData(const CNode& peer, const uint256& txid,
 
     {
         LOCK(cs_main);
-
         // Otherwise, the transaction must have been announced recently.
-        if (State(peer.GetId())->m_recently_announced_invs.contains(txid)) {
+        if (State(peer.GetId())->m_recently_announced_invs.contains(txid_or_wtxid)) {
             // If it was, it can be relayed from either the mempool...
             if (txinfo.tx) return std::move(txinfo.tx);
             // ... or the relay pool.
-            auto mi = mapRelay.find(txid);
+            auto mi = mapRelay.find(txid_or_wtxid);
             if (mi != mapRelay.end()) return mi->second;
         }
     }
@@ -1676,7 +1719,7 @@ void static ProcessGetData(CNode& pfrom, const CChainParams& chainparams, CConnm
     // Process as many TX items from the front of the getdata queue as
     // possible, since they're common and it's efficient to batch process
     // them.
-    while (it != pfrom.vRecvGetData.end() && (it->type == MSG_TX || it->type == MSG_WITNESS_TX)) {
+    while (it != pfrom.vRecvGetData.end() && (it->type == MSG_TX || it->type == MSG_WITNESS_TX || it->type == MSG_WTX)) {
         if (interruptMsgProc) return;
         // The send buffer provides backpressure. If there's no space in
         // the buffer, pause processing until the next call.
@@ -1689,11 +1732,12 @@ void static ProcessGetData(CNode& pfrom, const CChainParams& chainparams, CConnm
             continue;
         }
 
-        CTransactionRef tx = FindTxForGetData(pfrom, inv.hash, mempool_req, now);
+        CTransactionRef tx = FindTxForGetData(pfrom, inv.hash, inv.type == MSG_WTX, mempool_req, now);
         if (tx) {
+            // WTX and WITNESS_TX imply we serialize with witness
             int nSendFlags = (inv.type == MSG_TX ? SERIALIZE_TRANSACTION_NO_WITNESS : 0);
             connman.PushMessage(&pfrom, msgMaker.Make(nSendFlags, NetMsgType::TX, *tx));
-            mempool.RemoveUnbroadcastTx(inv.hash);
+            mempool.RemoveUnbroadcastTx(tx->GetHash());
             // As we're going to send tx, make sure its unconfirmed parents are made requestable.
             for (const auto& txin : tx->vin) {
                 auto txinfo = mempool.info(txin.prevout.hash);
@@ -1972,7 +2016,7 @@ void static ProcessOrphanTx(CConnman& connman, CTxMemPool& mempool, std::set<uin
         if (setMisbehaving.count(fromPeer)) continue;
         if (AcceptToMemoryPool(mempool, orphan_state, porphanTx, &removed_txn, false /* bypass_limits */, 0 /* nAbsurdFee */)) {
             LogPrint(BCLog::MEMPOOL, "   accepted orphan tx %s\n", orphanHash.ToString());
-            RelayTransaction(orphanHash, connman);
+            RelayTransaction(orphanHash, porphanTx->GetWitnessHash(), connman);
             for (unsigned int i = 0; i < orphanTx.vout.size(); i++) {
                 auto it_by_prev = mapOrphanTransactionsByPrev.find(COutPoint(orphanHash, i));
                 if (it_by_prev != mapOrphanTransactionsByPrev.end()) {
@@ -1997,12 +2041,22 @@ void static ProcessOrphanTx(CConnman& connman, CTxMemPool& mempool, std::set<uin
             // Has inputs but not accepted to mempool
             // Probably non-standard or insufficient fee
             LogPrint(BCLog::MEMPOOL, "   removed orphan tx %s\n", orphanHash.ToString());
-            if (!orphanTx.HasWitness() && orphan_state.GetResult() != TxValidationResult::TX_WITNESS_MUTATED) {
-                // Do not use rejection cache for witness transactions or
-                // witness-stripped transactions, as they can have been malleated.
-                // See https://github.com/bitcoin/bitcoin/issues/8279 for details.
+            if (orphan_state.GetResult() != TxValidationResult::TX_WITNESS_STRIPPED) {
+                // We can add the wtxid of this transaction to our reject filter.
+                // Do not add txids of witness transactions or witness-stripped
+                // transactions to the filter, as they can have been malleated;
+                // adding such txids to the reject filter would potentially
+                // interfere with relay of valid transactions from peers that
+                // do not support wtxid-based relay. See
+                // https://github.com/bitcoin/bitcoin/issues/8279 for details.
+                // We can remove this restriction (and always add wtxids to
+                // the filter even for witness stripped transactions) once
+                // wtxid-based relay is broadly deployed.
+                // See also comments in https://github.com/bitcoin/bitcoin/pull/18044#discussion_r443419034
+                // for concerns around weakening security of unupgraded nodes
+                // if we start doing this too early.
                 assert(recentRejects);
-                recentRejects->insert(orphanHash);
+                recentRejects->insert(orphanTx.GetWitnessHash());
             }
             EraseOrphanTx(orphanHash);
             done = true;
@@ -2319,6 +2373,10 @@ void ProcessMessage(
         if (pfrom.fInbound)
             PushNodeVersion(pfrom, connman, GetAdjustedTime());
 
+        if (nVersion >= WTXID_RELAY_VERSION) {
+            connman.PushMessage(&pfrom, CNetMsgMaker(INIT_PROTO_VERSION).Make(NetMsgType::WTXIDRELAY));
+        }
+
         connman.PushMessage(&pfrom, CNetMsgMaker(INIT_PROTO_VERSION).Make(NetMsgType::VERACK));
 
         pfrom.nServices = nServices;
@@ -2455,6 +2513,25 @@ void ProcessMessage(
         return;
     }
 
+    // Feature negotiation of wtxidrelay should happen between VERSION and
+    // VERACK, to avoid relay problems from switching after a connection is up
+    if (msg_type == NetMsgType::WTXIDRELAY) {
+        if (pfrom.fSuccessfullyConnected) {
+            // Disconnect peers that send wtxidrelay message after VERACK; this
+            // must be negotiated between VERSION and VERACK.
+            pfrom.fDisconnect = true;
+            return;
+        }
+        if (pfrom.nVersion >= WTXID_RELAY_VERSION) {
+            LOCK(cs_main);
+            if (!State(pfrom.GetId())->m_wtxid_relay) {
+                State(pfrom.GetId())->m_wtxid_relay = true;
+                g_wtxid_relay_peers++;
+            }
+        }
+        return;
+    }
+
     if (!pfrom.fSuccessfullyConnected) {
         // Must have a verack message before anything else
         LOCK(cs_main);
@@ -2575,6 +2652,13 @@ void ProcessMessage(
             if (interruptMsgProc)
                 return;
 
+            // ignore INVs that don't match wtxidrelay setting
+            if (State(pfrom.GetId())->m_wtxid_relay) {
+                if (inv.type == MSG_TX) continue;
+            } else {
+                if (inv.type == MSG_WTX) continue;
+            }
+
             bool fAlreadyHave = AlreadyHave(inv, mempool);
             LogPrint(BCLog::NET, "got inv: %s  %s peer=%d\n", inv.ToString(), fAlreadyHave ? "have" : "new", pfrom.GetId());
 
@@ -2593,7 +2677,7 @@ void ProcessMessage(
                     best_block = &inv.hash;
                 }
             } else {
-                pfrom.AddInventoryKnown(inv);
+                pfrom.AddKnownTx(inv.hash);
                 if (fBlocksOnly) {
                     LogPrint(BCLog::NET, "transaction (%s) inv sent in violation of protocol, disconnecting peer=%d\n", inv.hash.ToString(), pfrom.GetId());
                     pfrom.fDisconnect = true;
@@ -2832,26 +2916,50 @@ void ProcessMessage(
         vRecv >> ptx;
         const CTransaction& tx = *ptx;
 
-        CInv inv(MSG_TX, tx.GetHash());
-        pfrom.AddInventoryKnown(inv);
+        const uint256& txid = ptx->GetHash();
+        const uint256& wtxid = ptx->GetWitnessHash();
 
         LOCK2(cs_main, g_cs_orphans);
 
+        CNodeState* nodestate = State(pfrom.GetId());
+
+        const uint256& hash = nodestate->m_wtxid_relay ? wtxid : txid;
+        pfrom.AddKnownTx(hash);
+        if (nodestate->m_wtxid_relay && txid != wtxid) {
+            // Insert txid into filterInventoryKnown, even for
+            // wtxidrelay peers. This prevents re-adding of
+            // unconfirmed parents to the recently_announced
+            // filter, when a child tx is requested. See
+            // ProcessGetData().
+            pfrom.AddKnownTx(txid);
+        }
+
         TxValidationState state;
 
-        CNodeState* nodestate = State(pfrom.GetId());
-        nodestate->m_tx_download.m_tx_announced.erase(inv.hash);
-        nodestate->m_tx_download.m_tx_in_flight.erase(inv.hash);
-        EraseTxRequest(inv.hash);
+        nodestate->m_tx_download.m_tx_announced.erase(hash);
+        nodestate->m_tx_download.m_tx_in_flight.erase(hash);
+        EraseTxRequest(hash);
 
         std::list<CTransactionRef> lRemovedTxn;
 
-        if (!AlreadyHave(inv, mempool) &&
+        // We do the AlreadyHave() check using wtxid, rather than txid - in the
+        // absence of witness malleation, this is strictly better, because the
+        // recent rejects filter may contain the wtxid but will never contain
+        // the txid of a segwit transaction that has been rejected.
+        // In the presence of witness malleation, it's possible that by only
+        // doing the check with wtxid, we could overlook a transaction which
+        // was confirmed with a different witness, or exists in our mempool
+        // with a different witness, but this has limited downside:
+        // mempool validation does its own lookup of whether we have the txid
+        // already; and an adversary can already relay us old transactions
+        // (older than our recency filter) if trying to DoS us, without any need
+        // for witness malleation.
+        if (!AlreadyHave(CInv(MSG_WTX, wtxid), mempool) &&
             AcceptToMemoryPool(mempool, state, ptx, &lRemovedTxn, false /* bypass_limits */, 0 /* nAbsurdFee */)) {
             mempool.check(&::ChainstateActive().CoinsTip());
-            RelayTransaction(tx.GetHash(), connman);
+            RelayTransaction(tx.GetHash(), tx.GetWitnessHash(), connman);
             for (unsigned int i = 0; i < tx.vout.size(); i++) {
-                auto it_by_prev = mapOrphanTransactionsByPrev.find(COutPoint(inv.hash, i));
+                auto it_by_prev = mapOrphanTransactionsByPrev.find(COutPoint(txid, i));
                 if (it_by_prev != mapOrphanTransactionsByPrev.end()) {
                     for (const auto& elem : it_by_prev->second) {
                         pfrom.orphan_work_set.insert(elem->first);
@@ -2882,10 +2990,17 @@ void ProcessMessage(
                 uint32_t nFetchFlags = GetFetchFlags(pfrom);
                 const auto current_time = GetTime<std::chrono::microseconds>();
 
-                for (const CTxIn& txin : tx.vin) {
-                    CInv _inv(MSG_TX | nFetchFlags, txin.prevout.hash);
-                    pfrom.AddInventoryKnown(_inv);
-                    if (!AlreadyHave(_inv, mempool)) RequestTx(State(pfrom.GetId()), _inv.hash, current_time);
+                if (!State(pfrom.GetId())->m_wtxid_relay) {
+                    for (const CTxIn& txin : tx.vin) {
+                        // Here, we only have the txid (and not wtxid) of the
+                        // inputs, so we only request parents from
+                        // non-wtxid-relay peers.
+                        // Eventually we should replace this with an improved
+                        // protocol for getting all unconfirmed parents.
+                        CInv _inv(MSG_TX | nFetchFlags, txin.prevout.hash);
+                        pfrom.AddKnownTx(txin.prevout.hash);
+                        if (!AlreadyHave(_inv, mempool)) RequestTx(State(pfrom.GetId()), _inv.hash, current_time);
+                    }
                 }
                 AddOrphanTx(ptx, pfrom.GetId());
 
@@ -2899,15 +3014,30 @@ void ProcessMessage(
                 LogPrint(BCLog::MEMPOOL, "not keeping orphan with rejected parents %s\n",tx.GetHash().ToString());
                 // We will continue to reject this tx since it has rejected
                 // parents so avoid re-requesting it from other peers.
+                // Here we add both the txid and the wtxid, as we know that
+                // regardless of what witness is provided, we will not accept
+                // this, so we don't need to allow for redownload of this txid
+                // from any of our non-wtxidrelay peers.
                 recentRejects->insert(tx.GetHash());
+                recentRejects->insert(tx.GetWitnessHash());
             }
         } else {
-            if (!tx.HasWitness() && state.GetResult() != TxValidationResult::TX_WITNESS_MUTATED) {
-                // Do not use rejection cache for witness transactions or
-                // witness-stripped transactions, as they can have been malleated.
-                // See https://github.com/bitcoin/bitcoin/issues/8279 for details.
+            if (state.GetResult() != TxValidationResult::TX_WITNESS_STRIPPED) {
+                // We can add the wtxid of this transaction to our reject filter.
+                // Do not add txids of witness transactions or witness-stripped
+                // transactions to the filter, as they can have been malleated;
+                // adding such txids to the reject filter would potentially
+                // interfere with relay of valid transactions from peers that
+                // do not support wtxid-based relay. See
+                // https://github.com/bitcoin/bitcoin/issues/8279 for details.
+                // We can remove this restriction (and always add wtxids to
+                // the filter even for witness stripped transactions) once
+                // wtxid-based relay is broadly deployed.
+                // See also comments in https://github.com/bitcoin/bitcoin/pull/18044#discussion_r443419034
+                // for concerns around weakening security of unupgraded nodes
+                // if we start doing this too early.
                 assert(recentRejects);
-                recentRejects->insert(tx.GetHash());
+                recentRejects->insert(tx.GetWitnessHash());
                 if (RecursiveDynamicUsage(*ptx) < 100000) {
                     AddToCompactExtraTransactions(ptx);
                 }
@@ -2924,7 +3054,7 @@ void ProcessMessage(
                     LogPrintf("Not relaying non-mempool transaction %s from forcerelay peer=%d\n", tx.GetHash().ToString(), pfrom.GetId());
                 } else {
                     LogPrintf("Force relaying tx %s from peer=%d\n", tx.GetHash().ToString(), pfrom.GetId());
-                    RelayTransaction(tx.GetHash(), connman);
+                    RelayTransaction(tx.GetHash(), tx.GetWitnessHash(), connman);
                 }
             }
         }
@@ -3564,7 +3694,7 @@ void ProcessMessage(
         vRecv >> vInv;
         if (vInv.size() <= MAX_PEER_TX_IN_FLIGHT + MAX_BLOCKS_IN_TRANSIT_PER_PEER) {
             for (CInv &inv : vInv) {
-                if (inv.type == MSG_TX || inv.type == MSG_WITNESS_TX) {
+                if (inv.type == MSG_TX || inv.type == MSG_WITNESS_TX || inv.type == MSG_WTX) {
                     // If we receive a NOTFOUND message for a txid we requested, erase
                     // it from our data structures for this peer.
                     auto in_flight_it = state->m_tx_download.m_tx_in_flight.find(inv.hash);
@@ -3849,17 +3979,19 @@ namespace {
 class CompareInvMempoolOrder
 {
     CTxMemPool *mp;
+    bool m_wtxid_relay;
 public:
-    explicit CompareInvMempoolOrder(CTxMemPool *_mempool)
+    explicit CompareInvMempoolOrder(CTxMemPool *_mempool, bool use_wtxid)
     {
         mp = _mempool;
+        m_wtxid_relay = use_wtxid;
     }
 
     bool operator()(std::set<uint256>::iterator a, std::set<uint256>::iterator b)
     {
         /* As std::make_heap produces a max-heap, we want the entries with the
          * fewest ancestors/highest fee to sort later. */
-        return mp->CompareDepthAndScore(*b, *a);
+        return mp->CompareDepthAndScore(*b, *a, m_wtxid_relay);
     }
 };
 }
@@ -4166,8 +4298,8 @@ bool PeerLogicValidation::SendMessages(CNode* pto)
                     LOCK(pto->m_tx_relay->cs_filter);
 
                     for (const auto& txinfo : vtxinfo) {
-                        const uint256& hash = txinfo.tx->GetHash();
-                        CInv inv(MSG_TX, hash);
+                        const uint256& hash = state.m_wtxid_relay ? txinfo.tx->GetWitnessHash() : txinfo.tx->GetHash();
+                        CInv inv(state.m_wtxid_relay ? MSG_WTX : MSG_TX, hash);
                         pto->m_tx_relay->setInventoryTxToSend.erase(hash);
                         // Don't send transactions that peers will not put into their mempool
                         if (txinfo.fee < filterrate.GetFee(txinfo.vsize)) {
@@ -4202,7 +4334,7 @@ bool PeerLogicValidation::SendMessages(CNode* pto)
                     }
                     // Topologically and fee-rate sort the inventory we send for privacy and priority reasons.
                     // A heap is used so that not all items need sorting if only a few are being sent.
-                    CompareInvMempoolOrder compareInvMempoolOrder(&m_mempool);
+                    CompareInvMempoolOrder compareInvMempoolOrder(&m_mempool, state.m_wtxid_relay);
                     std::make_heap(vInvTx.begin(), vInvTx.end(), compareInvMempoolOrder);
                     // No reason to drain out at many times the network's capacity,
                     // especially since we have many peers and some will draw much shorter delays.
@@ -4221,10 +4353,12 @@ bool PeerLogicValidation::SendMessages(CNode* pto)
                             continue;
                         }
                         // Not in the mempool anymore? don't bother sending it.
-                        auto txinfo = m_mempool.info(hash);
+                        auto txinfo = m_mempool.info(hash, state.m_wtxid_relay);
                         if (!txinfo.tx) {
                             continue;
                         }
+                        auto txid = txinfo.tx->GetHash();
+                        auto wtxid = txinfo.tx->GetWitnessHash();
                         // Peer told you to not send transactions at that feerate? Don't bother sending it.
                         if (txinfo.fee < filterrate.GetFee(txinfo.vsize)) {
                             continue;
@@ -4232,7 +4366,7 @@ bool PeerLogicValidation::SendMessages(CNode* pto)
                         if (pto->m_tx_relay->pfilter && !pto->m_tx_relay->pfilter->IsRelevantAndUpdate(*txinfo.tx)) continue;
                         // Send
                         State(pto->GetId())->m_recently_announced_invs.insert(hash);
-                        vInv.push_back(CInv(MSG_TX, hash));
+                        vInv.push_back(CInv(state.m_wtxid_relay ? MSG_WTX : MSG_TX, hash));
                         nRelayedTransactions++;
                         {
                             // Expire old relay messages
@@ -4242,9 +4376,14 @@ bool PeerLogicValidation::SendMessages(CNode* pto)
                                 vRelayExpiration.pop_front();
                             }
 
-                            auto ret = mapRelay.insert(std::make_pair(hash, std::move(txinfo.tx)));
+                            auto ret = mapRelay.emplace(txid, std::move(txinfo.tx));
                             if (ret.second) {
-                                vRelayExpiration.push_back(std::make_pair(nNow + std::chrono::microseconds{RELAY_TX_CACHE_TIME}.count(), ret.first));
+                                vRelayExpiration.emplace_back(nNow + std::chrono::microseconds{RELAY_TX_CACHE_TIME}.count(), ret.first);
+                            }
+                            // Add wtxid-based lookup into mapRelay as well, so that peers can request by wtxid
+                            auto ret2 = mapRelay.emplace(wtxid, ret.first->second);
+                            if (ret2.second) {
+                                vRelayExpiration.emplace_back(nNow + std::chrono::microseconds{RELAY_TX_CACHE_TIME}.count(), ret2.first);
                             }
                         }
                         if (vInv.size() == MAX_INV_SZ) {
@@ -4252,6 +4391,14 @@ bool PeerLogicValidation::SendMessages(CNode* pto)
                             vInv.clear();
                         }
                         pto->m_tx_relay->filterInventoryKnown.insert(hash);
+                        if (hash != txid) {
+                            // Insert txid into filterInventoryKnown, even for
+                            // wtxidrelay peers. This prevents re-adding of
+                            // unconfirmed parents to the recently_announced
+                            // filter, when a child tx is requested. See
+                            // ProcessGetData().
+                            pto->m_tx_relay->filterInventoryKnown.insert(txid);
+                        }
                     }
                 }
             }
@@ -4376,7 +4523,7 @@ bool PeerLogicValidation::SendMessages(CNode* pto)
             // Erase this entry from tx_process_time (it may be added back for
             // processing at a later time, see below)
             tx_process_time.erase(tx_process_time.begin());
-            CInv inv(MSG_TX | GetFetchFlags(*pto), txid);
+            CInv inv(state.m_wtxid_relay ? MSG_WTX : (MSG_TX | GetFetchFlags(*pto)), txid);
             if (!AlreadyHave(inv, m_mempool)) {
                 // If this transaction was last requested more than 1 minute ago,
                 // then request.
@@ -4395,7 +4542,15 @@ bool PeerLogicValidation::SendMessages(CNode* pto)
                     // up processing to happen after the download times out
                     // (with a slight delay for inbound peers, to prefer
                     // requests to outbound peers).
-                    const auto next_process_time = CalculateTxGetDataTime(txid, current_time, !state.fPreferredDownload);
+                    // Don't apply the txid-delay to re-requests of a
+                    // transaction; the heuristic of delaying requests to
+                    // txid-relay peers is to save bandwidth on initial
+                    // announcement of a transaction, and doesn't make sense
+                    // for a followup request if our first peer times out (and
+                    // would open us up to an attacker using inbound
+                    // wtxid-relay to prevent us from requesting transactions
+                    // from outbound txid-relay peers).
+                    const auto next_process_time = CalculateTxGetDataTime(txid, current_time, !state.fPreferredDownload, false);
                     tx_process_time.emplace(next_process_time, txid);
                 }
             } else {
@@ -4459,6 +4614,7 @@ public:
         // orphan transactions
         mapOrphanTransactions.clear();
         mapOrphanTransactionsByPrev.clear();
+        g_orphans_by_wtxid.clear();
     }
 };
 static CNetProcessingCleanup instance_of_cnetprocessingcleanup;
diff --git a/src/net_processing.h b/src/net_processing.h
index fa1555fbe6..0534828761 100644
--- a/src/net_processing.h
+++ b/src/net_processing.h
@@ -100,6 +100,6 @@ struct CNodeStateStats {
 bool GetNodeStateStats(NodeId nodeid, CNodeStateStats &stats);
 
 /** Relay transaction to every node */
-void RelayTransaction(const uint256&, const CConnman& connman);
+void RelayTransaction(const uint256& txid, const uint256& wtxid, const CConnman& connman) EXCLUSIVE_LOCKS_REQUIRED(cs_main);
 
 #endif // BITCOIN_NET_PROCESSING_H
diff --git a/src/node/transaction.cpp b/src/node/transaction.cpp
index 3841d8687d..5633abe817 100644
--- a/src/node/transaction.cpp
+++ b/src/node/transaction.cpp
@@ -80,9 +80,10 @@ TransactionError BroadcastTransaction(NodeContext& node, const CTransactionRef t
     if (relay) {
         // the mempool tracks locally submitted transactions to make a
         // best-effort of initial broadcast
-        node.mempool->AddUnbroadcastTx(hashTx);
+        node.mempool->AddUnbroadcastTx(hashTx, tx->GetWitnessHash());
 
-        RelayTransaction(hashTx, *node.connman);
+        LOCK(cs_main);
+        RelayTransaction(hashTx, tx->GetWitnessHash(), *node.connman);
     }
 
     return TransactionError::OK;
diff --git a/src/protocol.cpp b/src/protocol.cpp
index 2dfe4bee74..ee77ca3b94 100644
--- a/src/protocol.cpp
+++ b/src/protocol.cpp
@@ -46,6 +46,7 @@ const char *GETCFHEADERS="getcfheaders";
 const char *CFHEADERS="cfheaders";
 const char *GETCFCHECKPT="getcfcheckpt";
 const char *CFCHECKPT="cfcheckpt";
+const char *WTXIDRELAY="wtxidrelay";
 } // namespace NetMsgType
 
 /** All known message types. Keep this in the same order as the list of
@@ -83,6 +84,7 @@ const static std::string allNetMessageTypes[] = {
     NetMsgType::CFHEADERS,
     NetMsgType::GETCFCHECKPT,
     NetMsgType::CFCHECKPT,
+    NetMsgType::WTXIDRELAY,
 };
 const static std::vector<std::string> allNetMessageTypesVec(allNetMessageTypes, allNetMessageTypes+ARRAYLEN(allNetMessageTypes));
 
@@ -177,6 +179,8 @@ std::string CInv::GetCommand() const
     switch (masked)
     {
     case MSG_TX:             return cmd.append(NetMsgType::TX);
+    // WTX is not a message type, just an inv type
+    case MSG_WTX:            return cmd.append("wtx");
     case MSG_BLOCK:          return cmd.append(NetMsgType::BLOCK);
     case MSG_FILTERED_BLOCK: return cmd.append(NetMsgType::MERKLEBLOCK);
     case MSG_CMPCT_BLOCK:    return cmd.append(NetMsgType::CMPCTBLOCK);
diff --git a/src/protocol.h b/src/protocol.h
index 9ab63a30fb..d83da2034a 100644
--- a/src/protocol.h
+++ b/src/protocol.h
@@ -261,6 +261,12 @@ extern const char* GETCFCHECKPT;
  * evenly spaced filter headers for blocks on the requested chain.
  */
 extern const char* CFCHECKPT;
+/**
+ * Indicates that a node prefers to relay transactions via wtxid, rather than
+ * txid.
+ * @since protocol version 70016 as described by BIP 339.
+ */
+extern const char *WTXIDRELAY;
 }; // namespace NetMsgType
 
 /* Get a vector of all valid message types (see above) */
@@ -397,11 +403,12 @@ const uint32_t MSG_TYPE_MASK = 0xffffffff >> 2;
  * These numbers are defined by the protocol. When adding a new value, be sure
  * to mention it in the respective BIP.
  */
-enum GetDataMsg {
+enum GetDataMsg : uint32_t {
     UNDEFINED = 0,
     MSG_TX = 1,
     MSG_BLOCK = 2,
-    // The following can only occur in getdata. Invs always use TX or BLOCK.
+    MSG_WTX = 5,                                      //!< Defined in BIP 339
+    // The following can only occur in getdata. Invs always use TX/WTX or BLOCK.
     MSG_FILTERED_BLOCK = 3,                           //!< Defined in BIP37
     MSG_CMPCT_BLOCK = 4,                              //!< Defined in BIP152
     MSG_WITNESS_BLOCK = MSG_BLOCK | MSG_WITNESS_FLAG, //!< Defined in BIP144
diff --git a/src/txmempool.cpp b/src/txmempool.cpp
index 7d8eb8a323..1d9f6a4a46 100644
--- a/src/txmempool.cpp
+++ b/src/txmempool.cpp
@@ -726,12 +726,12 @@ void CTxMemPool::check(const CCoinsViewCache *pcoins) const
     assert(innerUsage == cachedInnerUsage);
 }
 
-bool CTxMemPool::CompareDepthAndScore(const uint256& hasha, const uint256& hashb)
+bool CTxMemPool::CompareDepthAndScore(const uint256& hasha, const uint256& hashb, bool wtxid)
 {
     LOCK(cs);
-    indexed_transaction_set::const_iterator i = mapTx.find(hasha);
+    indexed_transaction_set::const_iterator i = wtxid ? get_iter_from_wtxid(hasha) : mapTx.find(hasha);
     if (i == mapTx.end()) return false;
-    indexed_transaction_set::const_iterator j = mapTx.find(hashb);
+    indexed_transaction_set::const_iterator j = wtxid ? get_iter_from_wtxid(hashb) : mapTx.find(hashb);
     if (j == mapTx.end()) return true;
     uint64_t counta = i->GetCountWithAncestors();
     uint64_t countb = j->GetCountWithAncestors();
@@ -811,10 +811,10 @@ CTransactionRef CTxMemPool::get(const uint256& hash) const
     return i->GetSharedTx();
 }
 
-TxMempoolInfo CTxMemPool::info(const uint256& hash) const
+TxMempoolInfo CTxMemPool::info(const uint256& hash, bool wtxid) const
 {
     LOCK(cs);
-    indexed_transaction_set::const_iterator i = mapTx.find(hash);
+    indexed_transaction_set::const_iterator i = (wtxid ? get_iter_from_wtxid(hash) : mapTx.find(hash));
     if (i == mapTx.end())
         return TxMempoolInfo();
     return GetInfo(i);
@@ -917,8 +917,8 @@ bool CCoinsViewMemPool::GetCoin(const COutPoint &outpoint, Coin &coin) const {
 
 size_t CTxMemPool::DynamicMemoryUsage() const {
     LOCK(cs);
-    // Estimate the overhead of mapTx to be 12 pointers + an allocation, as no exact formula for boost::multi_index_contained is implemented.
-    return memusage::MallocUsage(sizeof(CTxMemPoolEntry) + 12 * sizeof(void*)) * mapTx.size() + memusage::DynamicUsage(mapNextTx) + memusage::DynamicUsage(mapDeltas) + memusage::DynamicUsage(mapLinks) + memusage::DynamicUsage(vTxHashes) + cachedInnerUsage;
+    // Estimate the overhead of mapTx to be 15 pointers + an allocation, as no exact formula for boost::multi_index_contained is implemented.
+    return memusage::MallocUsage(sizeof(CTxMemPoolEntry) + 15 * sizeof(void*)) * mapTx.size() + memusage::DynamicUsage(mapNextTx) + memusage::DynamicUsage(mapDeltas) + memusage::DynamicUsage(mapLinks) + memusage::DynamicUsage(vTxHashes) + cachedInnerUsage;
 }
 
 void CTxMemPool::RemoveUnbroadcastTx(const uint256& txid, const bool unchecked) {
diff --git a/src/txmempool.h b/src/txmempool.h
index 583f7614b7..d4e9845942 100644
--- a/src/txmempool.h
+++ b/src/txmempool.h
@@ -198,6 +198,22 @@ struct mempoolentry_txid
     }
 };
 
+// extracts a transaction witness-hash from CTxMemPoolEntry or CTransactionRef
+struct mempoolentry_wtxid
+{
+    typedef uint256 result_type;
+    result_type operator() (const CTxMemPoolEntry &entry) const
+    {
+        return entry.GetTx().GetWitnessHash();
+    }
+
+    result_type operator() (const CTransactionRef& tx) const
+    {
+        return tx->GetWitnessHash();
+    }
+};
+
+
 /** \class CompareTxMemPoolEntryByDescendantScore
  *
  *  Sort an entry by max(score/size of entry's tx, score/size with all descendants).
@@ -318,6 +334,7 @@ public:
 struct descendant_score {};
 struct entry_time {};
 struct ancestor_score {};
+struct index_by_wtxid {};
 
 class CBlockPolicyEstimator;
 
@@ -383,8 +400,9 @@ public:
  *
  * CTxMemPool::mapTx, and CTxMemPoolEntry bookkeeping:
  *
- * mapTx is a boost::multi_index that sorts the mempool on 4 criteria:
- * - transaction hash
+ * mapTx is a boost::multi_index that sorts the mempool on 5 criteria:
+ * - transaction hash (txid)
+ * - witness-transaction hash (wtxid)
  * - descendant feerate [we use max(feerate of tx, feerate of tx with all descendants)]
  * - time in mempool
  * - ancestor feerate [we use min(feerate of tx, feerate of tx with all unconfirmed ancestors)]
@@ -469,6 +487,12 @@ public:
         boost::multi_index::indexed_by<
             // sorted by txid
             boost::multi_index::hashed_unique<mempoolentry_txid, SaltedTxidHasher>,
+            // sorted by wtxid
+            boost::multi_index::hashed_unique<
+                boost::multi_index::tag<index_by_wtxid>,
+                mempoolentry_wtxid,
+                SaltedTxidHasher
+            >,
             // sorted by fee rate
             boost::multi_index::ordered_non_unique<
                 boost::multi_index::tag<descendant_score>,
@@ -549,8 +573,11 @@ private:
 
     std::vector<indexed_transaction_set::const_iterator> GetSortedDepthAndScore() const EXCLUSIVE_LOCKS_REQUIRED(cs);
 
-    /** track locally submitted transactions to periodically retry initial broadcast */
-    std::set<uint256> m_unbroadcast_txids GUARDED_BY(cs);
+    /**
+     * track locally submitted transactions to periodically retry initial broadcast
+     * map of txid -> wtxid
+     */
+    std::map<uint256, uint256> m_unbroadcast_txids GUARDED_BY(cs);
 
 public:
     indirectmap<COutPoint, const CTransaction*> mapNextTx GUARDED_BY(cs);
@@ -586,7 +613,7 @@ public:
 
     void clear();
     void _clear() EXCLUSIVE_LOCKS_REQUIRED(cs); //lock free
-    bool CompareDepthAndScore(const uint256& hasha, const uint256& hashb);
+    bool CompareDepthAndScore(const uint256& hasha, const uint256& hashb, bool wtxid=false);
     void queryHashes(std::vector<uint256>& vtxid) const;
     bool isSpent(const COutPoint& outpoint) const;
     unsigned int GetTransactionsUpdated() const;
@@ -689,24 +716,32 @@ public:
         return totalTxSize;
     }
 
-    bool exists(const uint256& hash) const
+    bool exists(const uint256& hash, bool wtxid=false) const
     {
         LOCK(cs);
+        if (wtxid) {
+            return (mapTx.get<index_by_wtxid>().count(hash) != 0);
+        }
         return (mapTx.count(hash) != 0);
     }
 
     CTransactionRef get(const uint256& hash) const;
-    TxMempoolInfo info(const uint256& hash) const;
+    txiter get_iter_from_wtxid(const uint256& wtxid) const EXCLUSIVE_LOCKS_REQUIRED(cs)
+    {
+        AssertLockHeld(cs);
+        return mapTx.project<0>(mapTx.get<index_by_wtxid>().find(wtxid));
+    }
+    TxMempoolInfo info(const uint256& hash, bool wtxid=false) const;
     std::vector<TxMempoolInfo> infoAll() const;
 
     size_t DynamicMemoryUsage() const;
 
     /** Adds a transaction to the unbroadcast set */
-    void AddUnbroadcastTx(const uint256& txid) {
+    void AddUnbroadcastTx(const uint256& txid, const uint256& wtxid) {
         LOCK(cs);
         // Sanity Check: the transaction should also be in the mempool
         if (exists(txid)) {
-            m_unbroadcast_txids.insert(txid);
+            m_unbroadcast_txids[txid] = wtxid;
         }
     }
 
@@ -714,7 +749,7 @@ public:
     void RemoveUnbroadcastTx(const uint256& txid, const bool unchecked = false);
 
     /** Returns transactions in unbroadcast set */
-    std::set<uint256> GetUnbroadcastTxs() const {
+    std::map<uint256, uint256> GetUnbroadcastTxs() const {
         LOCK(cs);
         return m_unbroadcast_txids;
     }
diff --git a/src/validation.cpp b/src/validation.cpp
index a9cb99aafb..5aa3d315d5 100644
--- a/src/validation.cpp
+++ b/src/validation.cpp
@@ -939,7 +939,7 @@ bool MemPoolAccept::PolicyScriptChecks(ATMPArgs& args, Workspace& ws, Precompute
         if (!tx.HasWitness() && CheckInputScripts(tx, state_dummy, m_view, scriptVerifyFlags & ~(SCRIPT_VERIFY_WITNESS | SCRIPT_VERIFY_CLEANSTACK), true, false, txdata) &&
                 !CheckInputScripts(tx, state_dummy, m_view, scriptVerifyFlags & ~SCRIPT_VERIFY_CLEANSTACK, true, false, txdata)) {
             // Only the witness is missing, so the transaction itself may be fine.
-            state.Invalid(TxValidationResult::TX_WITNESS_MUTATED,
+            state.Invalid(TxValidationResult::TX_WITNESS_STRIPPED,
                     state.GetRejectReason(), state.GetDebugMessage());
         }
         return false; // state filled in by CheckInputScripts
@@ -5084,19 +5084,22 @@ bool LoadMempool(CTxMemPool& pool)
         }
 
         // TODO: remove this try except in v0.22
+        std::map<uint256, uint256> unbroadcast_txids;
         try {
-          std::set<uint256> unbroadcast_txids;
           file >> unbroadcast_txids;
           unbroadcast = unbroadcast_txids.size();
-
-          for (const auto& txid : unbroadcast_txids) {
-            pool.AddUnbroadcastTx(txid);
-          }
         } catch (const std::exception&) {
           // mempool.dat files created prior to v0.21 will not have an
           // unbroadcast set. No need to log a failure if parsing fails here.
         }
-
+        for (const auto& elem : unbroadcast_txids) {
+            // Don't add unbroadcast transactions that didn't get back into the
+            // mempool.
+            const CTransactionRef& added_tx = pool.get(elem.first);
+            if (added_tx != nullptr) {
+                pool.AddUnbroadcastTx(elem.first, added_tx->GetWitnessHash());
+            }
+        }
     } catch (const std::exception& e) {
         LogPrintf("Failed to deserialize mempool data on disk: %s. Continuing anyway.\n", e.what());
         return false;
@@ -5112,7 +5115,7 @@ bool DumpMempool(const CTxMemPool& pool)
 
     std::map<uint256, CAmount> mapDeltas;
     std::vector<TxMempoolInfo> vinfo;
-    std::set<uint256> unbroadcast_txids;
+    std::map<uint256, uint256> unbroadcast_txids;
 
     static Mutex dump_mutex;
     LOCK(dump_mutex);
diff --git a/src/version.h b/src/version.h
index e5d1f5a7f9..b5f379e1b8 100644
--- a/src/version.h
+++ b/src/version.h
@@ -9,7 +9,7 @@
  * network protocol versioning
  */
 
-static const int PROTOCOL_VERSION = 70015;
+static const int PROTOCOL_VERSION = 70016;
 
 //! initial proto version, to be increased after version/verack negotiation
 static const int INIT_PROTO_VERSION = 209;
@@ -35,4 +35,7 @@ static const int SHORT_IDS_BLOCKS_VERSION = 70014;
 //! not banning for invalid compact blocks starts with this version
 static const int INVALID_CB_NO_BAN_VERSION = 70015;
 
+//! "wtxidrelay" command for wtxid-based relay starts with this version
+static const int WTXID_RELAY_VERSION = 70016;
+
 #endif // BITCOIN_VERSION_H
diff --git a/test/functional/mempool_packages.py b/test/functional/mempool_packages.py
index 5b7216b253..542d24f4be 100755
--- a/test/functional/mempool_packages.py
+++ b/test/functional/mempool_packages.py
@@ -69,14 +69,19 @@ class MempoolPackagesTest(BitcoinTestFramework):
         fee = Decimal("0.0001")
         # MAX_ANCESTORS transactions off a confirmed tx should be fine
         chain = []
+        witness_chain = []
         for i in range(MAX_ANCESTORS):
             (txid, sent_value) = self.chain_transaction(self.nodes[0], txid, 0, value, fee, 1)
             value = sent_value
             chain.append(txid)
+            # We need the wtxids to check P2P announcements
+            fulltx = self.nodes[0].getrawtransaction(txid)
+            witnesstx = self.nodes[0].decoderawtransaction(fulltx, True)
+            witness_chain.append(witnesstx['hash'])
 
         # Wait until mempool transactions have passed initial broadcast (sent inv and received getdata)
         # Otherwise, getrawmempool may be inconsistent with getmempoolentry if unbroadcast changes in between
-        self.nodes[0].p2p.wait_for_broadcast(chain)
+        self.nodes[0].p2p.wait_for_broadcast(witness_chain)
 
         # Check mempool has MAX_ANCESTORS transactions in it, and descendant and ancestor
         # count and fees should look correct
diff --git a/test/functional/p2p_blocksonly.py b/test/functional/p2p_blocksonly.py
index f42a343042..27e6b669f6 100755
--- a/test/functional/p2p_blocksonly.py
+++ b/test/functional/p2p_blocksonly.py
@@ -52,7 +52,7 @@ class P2PBlocksOnly(BitcoinTestFramework):
         self.log.info('Check that txs from rpc are not rejected and relayed to other peers')
         assert_equal(self.nodes[0].getpeerinfo()[0]['relaytxes'], True)
         txid = self.nodes[0].testmempoolaccept([sigtx])[0]['txid']
-        with self.nodes[0].assert_debug_log(['received getdata for: tx {} peer=1'.format(txid)]):
+        with self.nodes[0].assert_debug_log(['received getdata for: wtx {} peer=1'.format(txid)]):
             self.nodes[0].sendrawtransaction(sigtx)
             self.nodes[0].p2p.wait_for_tx(txid)
             assert_equal(self.nodes[0].getmempoolinfo()['size'], 1)
diff --git a/test/functional/p2p_feefilter.py b/test/functional/p2p_feefilter.py
index f939ea965c..73afe9adc4 100755
--- a/test/functional/p2p_feefilter.py
+++ b/test/functional/p2p_feefilter.py
@@ -7,7 +7,7 @@
 from decimal import Decimal
 import time
 
-from test_framework.messages import MSG_TX, msg_feefilter
+from test_framework.messages import MSG_TX, MSG_WTX, msg_feefilter
 from test_framework.mininode import mininode_lock, P2PInterface
 from test_framework.test_framework import BitcoinTestFramework
 from test_framework.util import assert_equal
@@ -45,7 +45,7 @@ class TestP2PConn(P2PInterface):
 
     def on_inv(self, message):
         for i in message.inv:
-            if (i.type == MSG_TX):
+            if (i.type == MSG_TX) or (i.type == MSG_WTX):
                 self.txinvs.append(hashToHex(i.hash))
 
     def clear_invs(self):
diff --git a/test/functional/p2p_segwit.py b/test/functional/p2p_segwit.py
index 25dd765442..9915b844d1 100755
--- a/test/functional/p2p_segwit.py
+++ b/test/functional/p2p_segwit.py
@@ -25,6 +25,7 @@ from test_framework.messages import (
     MSG_BLOCK,
     MSG_TX,
     MSG_WITNESS_FLAG,
+    MSG_WTX,
     NODE_NETWORK,
     NODE_WITNESS,
     msg_no_witness_block,
@@ -34,6 +35,7 @@ from test_framework.messages import (
     msg_tx,
     msg_block,
     msg_no_witness_tx,
+    msg_verack,
     ser_uint256,
     ser_vector,
     sha256,
@@ -81,6 +83,7 @@ from test_framework.util import (
     softfork_active,
     hex_str_to_bytes,
     assert_raises_rpc_error,
+    wait_until,
 )
 
 # The versionbit bit used to signal activation of SegWit
@@ -143,25 +146,47 @@ def test_witness_block(node, p2p, block, accepted, with_witness=True, reason=Non
 
 
 class TestP2PConn(P2PInterface):
-    def __init__(self):
+    def __init__(self, wtxidrelay=False):
         super().__init__()
         self.getdataset = set()
+        self.last_wtxidrelay = []
+        self.lastgetdata = []
+        self.wtxidrelay = wtxidrelay
 
     # Avoid sending out msg_getdata in the mininode thread as a reply to invs.
     # They are not needed and would only lead to races because we send msg_getdata out in the test thread
     def on_inv(self, message):
         pass
 
+    def on_version(self, message):
+        if self.wtxidrelay:
+            super().on_version(message)
+        else:
+            self.send_message(msg_verack())
+            self.nServices = message.nServices
+
     def on_getdata(self, message):
+        self.lastgetdata = message.inv
         for inv in message.inv:
             self.getdataset.add(inv.hash)
 
-    def announce_tx_and_wait_for_getdata(self, tx, timeout=60, success=True):
+    def on_wtxidrelay(self, message):
+        self.last_wtxidrelay.append(message)
+
+    def announce_tx_and_wait_for_getdata(self, tx, timeout=60, success=True, use_wtxid=False):
         with mininode_lock:
             self.last_message.pop("getdata", None)
-        self.send_message(msg_inv(inv=[CInv(MSG_TX, tx.sha256)]))
+        if use_wtxid:
+            wtxid = tx.calc_sha256(True)
+            self.send_message(msg_inv(inv=[CInv(MSG_WTX, wtxid)]))
+        else:
+            self.send_message(msg_inv(inv=[CInv(MSG_TX, tx.sha256)]))
+
         if success:
-            self.wait_for_getdata([tx.sha256], timeout)
+            if use_wtxid:
+                self.wait_for_getdata([wtxid], timeout)
+            else:
+                self.wait_for_getdata([tx.sha256], timeout)
         else:
             time.sleep(timeout)
             assert not self.last_message.get("getdata")
@@ -277,6 +302,7 @@ class SegWitTest(BitcoinTestFramework):
         self.test_upgrade_after_activation()
         self.test_witness_sigops()
         self.test_superfluous_witness()
+        self.test_wtxid_relay()
 
     # Individual tests
 
@@ -1270,7 +1296,6 @@ class SegWitTest(BitcoinTestFramework):
         test_transaction_acceptance(self.nodes[0], self.test_node, tx, with_witness=True, accepted=False)
 
         # Verify that removing the witness succeeds.
-        self.test_node.announce_tx_and_wait_for_getdata(tx)
         test_transaction_acceptance(self.nodes[0], self.test_node, tx, with_witness=False, accepted=True)
 
         # Now try to add extra witness data to a valid witness tx.
@@ -1297,8 +1322,6 @@ class SegWitTest(BitcoinTestFramework):
         # Node will not be blinded to the transaction
         self.std_node.announce_tx_and_wait_for_getdata(tx3)
         test_transaction_acceptance(self.nodes[1], self.std_node, tx3, True, False, 'tx-size')
-        self.std_node.announce_tx_and_wait_for_getdata(tx3)
-        test_transaction_acceptance(self.nodes[1], self.std_node, tx3, True, False, 'tx-size')
 
         # Remove witness stuffing, instead add extra witness push on stack
         tx3.vout[0] = CTxOut(tx2.vout[0].nValue - 1000, CScript([OP_TRUE, OP_DROP] * 15 + [OP_TRUE]))
@@ -2016,6 +2039,11 @@ class SegWitTest(BitcoinTestFramework):
 
         # TODO: test p2sh sigop counting
 
+        # Cleanup and prep for next test
+        self.utxo.pop(0)
+        self.utxo.append(UTXO(tx2.sha256, 0, tx2.vout[0].nValue))
+
+    @subtest  # type: ignore
     def test_superfluous_witness(self):
         # Serialization of tx that puts witness flag to 3 always
         def serialize_with_bogus_witness(tx):
@@ -2059,6 +2087,67 @@ class SegWitTest(BitcoinTestFramework):
         with self.nodes[0].assert_debug_log(['Unknown transaction optional data']):
             self.nodes[0].p2p.send_and_ping(msg_bogus_tx(tx))
 
+    @subtest  # type: ignore
+    def test_wtxid_relay(self):
+        # Use brand new nodes to avoid contamination from earlier tests
+        self.wtx_node = self.nodes[0].add_p2p_connection(TestP2PConn(wtxidrelay=True), services=NODE_NETWORK | NODE_WITNESS)
+        self.tx_node = self.nodes[0].add_p2p_connection(TestP2PConn(wtxidrelay=False), services=NODE_NETWORK | NODE_WITNESS)
+
+        # Check wtxidrelay feature negotiation message through connecting a new peer
+        def received_wtxidrelay():
+            return (len(self.wtx_node.last_wtxidrelay) > 0)
+        wait_until(received_wtxidrelay, timeout=60, lock=mininode_lock)
+
+        # Create a Segwit output from the latest UTXO
+        # and announce it to the network
+        witness_program = CScript([OP_TRUE])
+        witness_hash = sha256(witness_program)
+        script_pubkey = CScript([OP_0, witness_hash])
+
+        tx = CTransaction()
+        tx.vin.append(CTxIn(COutPoint(self.utxo[0].sha256, self.utxo[0].n), b""))
+        tx.vout.append(CTxOut(self.utxo[0].nValue - 1000, script_pubkey))
+        tx.rehash()
+
+        # Create a Segwit transaction
+        tx2 = CTransaction()
+        tx2.vin.append(CTxIn(COutPoint(tx.sha256, 0), b""))
+        tx2.vout.append(CTxOut(tx.vout[0].nValue - 1000, script_pubkey))
+        tx2.wit.vtxinwit.append(CTxInWitness())
+        tx2.wit.vtxinwit[0].scriptWitness.stack = [witness_program]
+        tx2.rehash()
+
+        # Announce Segwit transaction with wtxid
+        # and wait for getdata
+        self.wtx_node.announce_tx_and_wait_for_getdata(tx2, use_wtxid=True)
+        with mininode_lock:
+            lgd = self.wtx_node.lastgetdata[:]
+        assert_equal(lgd, [CInv(MSG_WTX, tx2.calc_sha256(True))])
+
+        # Announce Segwit transaction from non wtxidrelay peer
+        # and wait for getdata
+        self.tx_node.announce_tx_and_wait_for_getdata(tx2, use_wtxid=False)
+        with mininode_lock:
+            lgd = self.tx_node.lastgetdata[:]
+        assert_equal(lgd, [CInv(MSG_TX|MSG_WITNESS_FLAG, tx2.sha256)])
+
+        # Send tx2 through; it's an orphan so won't be accepted
+        with mininode_lock:
+            self.tx_node.last_message.pop("getdata", None)
+        test_transaction_acceptance(self.nodes[0], self.tx_node, tx2, with_witness=True, accepted=False)
+
+        # Expect a request for parent (tx) due to use of non-WTX peer
+        self.tx_node.wait_for_getdata([tx.sha256], 60)
+        with mininode_lock:
+            lgd = self.tx_node.lastgetdata[:]
+        assert_equal(lgd, [CInv(MSG_TX|MSG_WITNESS_FLAG, tx.sha256)])
+
+        # Send tx through
+        test_transaction_acceptance(self.nodes[0], self.tx_node, tx, with_witness=False, accepted=True)
+
+        # Check tx2 is there now
+        assert_equal(tx2.hash in self.nodes[0].getrawmempool(), True)
+
 
 if __name__ == '__main__':
     SegWitTest().main()
diff --git a/test/functional/p2p_tx_download.py b/test/functional/p2p_tx_download.py
index 10f5eea0e5..2527edc135 100755
--- a/test/functional/p2p_tx_download.py
+++ b/test/functional/p2p_tx_download.py
@@ -12,6 +12,7 @@ from test_framework.messages import (
     FromHex,
     MSG_TX,
     MSG_TYPE_MASK,
+    MSG_WTX,
     msg_inv,
     msg_notfound,
 )
@@ -36,7 +37,7 @@ class TestP2PConn(P2PInterface):
 
     def on_getdata(self, message):
         for i in message.inv:
-            if i.type & MSG_TYPE_MASK == MSG_TX:
+            if i.type & MSG_TYPE_MASK == MSG_TX or i.type & MSG_TYPE_MASK == MSG_WTX:
                 self.tx_getdata_count += 1
 
 
@@ -44,12 +45,13 @@ class TestP2PConn(P2PInterface):
 GETDATA_TX_INTERVAL = 60  # seconds
 MAX_GETDATA_RANDOM_DELAY = 2  # seconds
 INBOUND_PEER_TX_DELAY = 2  # seconds
+TXID_RELAY_DELAY = 2 # seconds
 MAX_GETDATA_IN_FLIGHT = 100
 TX_EXPIRY_INTERVAL = GETDATA_TX_INTERVAL * 10
 
 # Python test constants
 NUM_INBOUND = 10
-MAX_GETDATA_INBOUND_WAIT = GETDATA_TX_INTERVAL + MAX_GETDATA_RANDOM_DELAY + INBOUND_PEER_TX_DELAY
+MAX_GETDATA_INBOUND_WAIT = GETDATA_TX_INTERVAL + MAX_GETDATA_RANDOM_DELAY + INBOUND_PEER_TX_DELAY + TXID_RELAY_DELAY
 
 
 class TxDownloadTest(BitcoinTestFramework):
@@ -63,7 +65,7 @@ class TxDownloadTest(BitcoinTestFramework):
         txid = 0xdeadbeef
 
         self.log.info("Announce the txid from each incoming peer to node 0")
-        msg = msg_inv([CInv(t=MSG_TX, h=txid)])
+        msg = msg_inv([CInv(t=MSG_WTX, h=txid)])
         for p in self.nodes[0].p2ps:
             p.send_and_ping(msg)
 
@@ -135,13 +137,13 @@ class TxDownloadTest(BitcoinTestFramework):
         with mininode_lock:
             p.tx_getdata_count = 0
 
-        p.send_message(msg_inv([CInv(t=MSG_TX, h=i) for i in txids]))
+        p.send_message(msg_inv([CInv(t=MSG_WTX, h=i) for i in txids]))
         wait_until(lambda: p.tx_getdata_count >= MAX_GETDATA_IN_FLIGHT, lock=mininode_lock)
         with mininode_lock:
             assert_equal(p.tx_getdata_count, MAX_GETDATA_IN_FLIGHT)
 
         self.log.info("Now check that if we send a NOTFOUND for a transaction, we'll get one more request")
-        p.send_message(msg_notfound(vec=[CInv(t=MSG_TX, h=txids[0])]))
+        p.send_message(msg_notfound(vec=[CInv(t=MSG_WTX, h=txids[0])]))
         wait_until(lambda: p.tx_getdata_count >= MAX_GETDATA_IN_FLIGHT + 1, timeout=10, lock=mininode_lock)
         with mininode_lock:
             assert_equal(p.tx_getdata_count, MAX_GETDATA_IN_FLIGHT + 1)
diff --git a/test/functional/test_framework/messages.py b/test/functional/test_framework/messages.py
index 12302cdbc3..2462a9a6db 100755
--- a/test/functional/test_framework/messages.py
+++ b/test/functional/test_framework/messages.py
@@ -31,7 +31,7 @@ from test_framework.siphash import siphash256
 from test_framework.util import hex_str_to_bytes, assert_equal
 
 MIN_VERSION_SUPPORTED = 60001
-MY_VERSION = 70014  # past bip-31 for ping/pong
+MY_VERSION = 70016  # past wtxid relay
 MY_SUBVERSION = b"/python-mininode-tester:0.0.3/"
 MY_RELAY = 1 # from version 70001 onwards, fRelay should be appended to version messages (BIP37)
 
@@ -59,6 +59,7 @@ MSG_TX = 1
 MSG_BLOCK = 2
 MSG_FILTERED_BLOCK = 3
 MSG_CMPCT_BLOCK = 4
+MSG_WTX = 5
 MSG_WITNESS_FLAG = 1 << 30
 MSG_TYPE_MASK = 0xffffffff >> 2
 
@@ -242,7 +243,8 @@ class CInv:
         MSG_TX | MSG_WITNESS_FLAG: "WitnessTx",
         MSG_BLOCK | MSG_WITNESS_FLAG: "WitnessBlock",
         MSG_FILTERED_BLOCK: "filtered Block",
-        4: "CompactBlock"
+        4: "CompactBlock",
+        5: "WTX",
     }
 
     def __init__(self, t=0, h=0):
@@ -263,6 +265,9 @@ class CInv:
         return "CInv(type=%s hash=%064x)" \
             % (self.typemap[self.type], self.hash)
 
+    def __eq__(self, other):
+        return isinstance(other, CInv) and self.hash == other.hash and self.type == other.type
+
 
 class CBlockLocator:
     __slots__ = ("nVersion", "vHave")
@@ -1124,6 +1129,22 @@ class msg_tx:
     def __repr__(self):
         return "msg_tx(tx=%s)" % (repr(self.tx))
 
+class msg_wtxidrelay:
+    __slots__ = ()
+    msgtype = b"wtxidrelay"
+
+    def __init__(self):
+        pass
+
+    def deserialize(self, f):
+        pass
+
+    def serialize(self):
+        return b""
+
+    def __repr__(self):
+        return "msg_wtxidrelay()"
+
 
 class msg_no_witness_tx(msg_tx):
     __slots__ = ()
diff --git a/test/functional/test_framework/mininode.py b/test/functional/test_framework/mininode.py
index e6da33763d..f68c1a9ddd 100755
--- a/test/functional/test_framework/mininode.py
+++ b/test/functional/test_framework/mininode.py
@@ -59,6 +59,8 @@ from test_framework.messages import (
     MSG_TYPE_MASK,
     msg_verack,
     msg_version,
+    MSG_WTX,
+    msg_wtxidrelay,
     NODE_NETWORK,
     NODE_WITNESS,
     sha256,
@@ -96,6 +98,7 @@ MESSAGEMAP = {
     b"tx": msg_tx,
     b"verack": msg_verack,
     b"version": msg_version,
+    b"wtxidrelay": msg_wtxidrelay,
 }
 
 MAGIC_BYTES = {
@@ -356,6 +359,7 @@ class P2PInterface(P2PConnection):
     def on_sendcmpct(self, message): pass
     def on_sendheaders(self, message): pass
     def on_tx(self, message): pass
+    def on_wtxidrelay(self, message): pass
 
     def on_inv(self, message):
         want = msg_getdata()
@@ -373,6 +377,8 @@ class P2PInterface(P2PConnection):
 
     def on_version(self, message):
         assert message.nVersion >= MIN_VERSION_SUPPORTED, "Version {} received. Test framework only supports versions greater than {}".format(message.nVersion, MIN_VERSION_SUPPORTED)
+        if message.nVersion >= 70016:
+            self.send_message(msg_wtxidrelay())
         self.send_message(msg_verack())
         self.nServices = message.nServices
 
@@ -654,7 +660,7 @@ class P2PTxInvStore(P2PInterface):
         super().on_inv(message) # Send getdata in response.
         # Store how many times invs have been received for each tx.
         for i in message.inv:
-            if i.type == MSG_TX:
+            if (i.type == MSG_TX) or (i.type == MSG_WTX):
                 # save txid
                 self.tx_invs_received[i.hash] += 1