[policy] explicitly require non-v3 for CPFP carve out

This carve out is intended to allow a second child under restricted
circumstances, but this topology is not allowed for v3 transactions.

As CPFP carve out does not explicitly require a second child to actually
exist, it has the effect of granting a free +10KvB descendant size limit
when a single child is enough to bust the descendant limit.

Author: glozow

src/validation.cpp

@@ -953,7 +953,8 @@ bool MemPoolAccept::PreChecks(ATMPArgs& args, Workspace& ws)
         // If the new transaction is relatively small (up to 40k weight)
         // and has at most one ancestor (ie ancestor limit of 2, including
         // the new transaction), allow it if its parent has exactly the
-        // descendant limit descendants.
+        // descendant limit descendants. The transaction also cannot be v3,
+        // as its topology restrictions do not allow a second child.
         //
         // This allows protocols which rely on distrusting counterparties
         // being able to broadcast descendants of an unconfirmed transaction
@@ -967,7 +968,7 @@ bool MemPoolAccept::PreChecks(ATMPArgs& args, Workspace& ws)
             .descendant_size_vbytes = maybe_rbf_limits.descendant_size_vbytes + EXTRA_DESCENDANT_TX_SIZE_LIMIT,
         };
         const auto error_message{util::ErrorString(ancestors).original};
-        if (ws.m_vsize > EXTRA_DESCENDANT_TX_SIZE_LIMIT) {
+        if (ws.m_vsize > EXTRA_DESCENDANT_TX_SIZE_LIMIT || ws.m_ptx->nVersion == 3) {
             return state.Invalid(TxValidationResult::TX_MEMPOOL_POLICY, "too-long-mempool-chain", error_message);
         }
         ancestors = m_pool.CalculateMemPoolAncestors(*entry, cpfp_carve_out_limits);