net: CNetAddr: add support to (un)serialize as ADDRv2

Co-authored-by: Carl Dong <[email protected]>

Author: vasild

src/netaddress.cpp

@@ -14,10 +14,65 @@
 #include <algorithm>
 #include <array>
 #include <cstdint>
+#include <ios>
 #include <iterator>
 #include <tuple>
 
 constexpr size_t CNetAddr::V1_SERIALIZATION_SIZE;
+constexpr size_t CNetAddr::MAX_ADDRV2_SIZE;
+
+CNetAddr::BIP155Network CNetAddr::GetBIP155Network() const
+{
+    switch (m_net) {
+    case NET_IPV4:
+        return BIP155Network::IPV4;
+    case NET_IPV6:
+        return BIP155Network::IPV6;
+    case NET_ONION:
+        return BIP155Network::TORV2;
+    case NET_INTERNAL:   // should have been handled before calling this function
+    case NET_UNROUTABLE: // m_net is never and should not be set to NET_UNROUTABLE
+    case NET_MAX:        // m_net is never and should not be set to NET_MAX
+        assert(false);
+    } // no default case, so the compiler can warn about missing cases
+
+    assert(false);
+}
+
+bool CNetAddr::SetNetFromBIP155Network(uint8_t possible_bip155_net, size_t address_size)
+{
+    switch (possible_bip155_net) {
+    case BIP155Network::IPV4:
+        if (address_size == ADDR_IPV4_SIZE) {
+            m_net = NET_IPV4;
+            return true;
+        }
+        throw std::ios_base::failure(
+            strprintf("BIP155 IPv4 address with length %u (should be %u)", address_size,
+                      ADDR_IPV4_SIZE));
+    case BIP155Network::IPV6:
+        if (address_size == ADDR_IPV6_SIZE) {
+            m_net = NET_IPV6;
+            return true;
+        }
+        throw std::ios_base::failure(
+            strprintf("BIP155 IPv6 address with length %u (should be %u)", address_size,
+                      ADDR_IPV6_SIZE));
+    case BIP155Network::TORV2:
+        if (address_size == ADDR_TORV2_SIZE) {
+            m_net = NET_ONION;
+            return true;
+        }
+        throw std::ios_base::failure(
+            strprintf("BIP155 TORv2 address with length %u (should be %u)", address_size,
+                      ADDR_TORV2_SIZE));
+    }
+
+    // Don't throw on addresses with unknown network ids (maybe from the future).
+    // Instead silently drop them and have the unserialization code consume
+    // subsequent ones which may be known to us.
+    return false;
+}
 
 /**
  * Construct an unspecified IPv6 network address (::/128).

src/netaddress.h

@@ -13,12 +13,24 @@
 #include <compat.h>
 #include <prevector.h>
 #include <serialize.h>
+#include <tinyformat.h>
+#include <util/strencodings.h>
+#include <util/string.h>
 
 #include <array>
 #include <cstdint>
+#include <ios>
 #include <string>
 #include <vector>
 
+/**
+ * A flag that is ORed into the protocol version to designate that addresses
+ * should be serialized in (unserialized from) v2 format (BIP155).
+ * Make sure that this does not collide with any of the values in `version.h`
+ * or with `SERIALIZE_TRANSACTION_NO_WITNESS`.
+ */
+static const int ADDRV2_FORMAT = 0x20000000;
+
 /**
  * A network type.
  * @note An address may belong to more than one network, for example `10.0.0.1`
@@ -177,7 +189,11 @@ class CNetAddr
         template <typename Stream>
         void Serialize(Stream& s) const
         {
-            SerializeV1Stream(s);
+            if (s.GetVersion() & ADDRV2_FORMAT) {
+                SerializeV2Stream(s);
+            } else {
+                SerializeV1Stream(s);
+            }
         }
 
         /**
@@ -186,17 +202,53 @@ class CNetAddr
         template <typename Stream>
         void Unserialize(Stream& s)
         {
-            UnserializeV1Stream(s);
+            if (s.GetVersion() & ADDRV2_FORMAT) {
+                UnserializeV2Stream(s);
+            } else {
+                UnserializeV1Stream(s);
+            }
         }
 
         friend class CSubNet;
 
     private:
+        /**
+         * BIP155 network ids recognized by this software.
+         */
+        enum BIP155Network : uint8_t {
+            IPV4 = 1,
+            IPV6 = 2,
+            TORV2 = 3,
+        };
+
         /**
          * Size of CNetAddr when serialized as ADDRv1 (pre-BIP155) (in bytes).
          */
         static constexpr size_t V1_SERIALIZATION_SIZE = ADDR_IPV6_SIZE;
 
+        /**
+         * Maximum size of an address as defined in BIP155 (in bytes).
+         * This is only the size of the address, not the entire CNetAddr object
+         * when serialized.
+         */
+        static constexpr size_t MAX_ADDRV2_SIZE = 512;
+
+        /**
+         * Get the BIP155 network id of this address.
+         * Must not be called for IsInternal() objects.
+         * @returns BIP155 network id
+         */
+        BIP155Network GetBIP155Network() const;
+
+        /**
+         * Set `m_net` from the provided BIP155 network id and size after validation.
+         * @retval true the network was recognized, is valid and `m_net` was set
+         * @retval false not recognised (from future?) and should be silently ignored
+         * @throws std::ios_base::failure if the network is one of the BIP155 founding
+         * networks recognized by this software (id 1..3) and has wrong address size.
+         */
+        bool SetNetFromBIP155Network(uint8_t possible_bip155_net, size_t address_size);
+
         /**
          * Serialize in pre-ADDRv2/BIP155 format to an array.
          * Some addresses (e.g. TORv3) cannot be serialized in pre-BIP155 format.
@@ -250,6 +302,25 @@ class CNetAddr
             s << serialized;
         }
 
+        /**
+         * Serialize as ADDRv2 / BIP155.
+         */
+        template <typename Stream>
+        void SerializeV2Stream(Stream& s) const
+        {
+            if (IsInternal()) {
+                // Serialize NET_INTERNAL as embedded in IPv6. We need to
+                // serialize such addresses from addrman.
+                s << static_cast<uint8_t>(BIP155Network::IPV6);
+                s << COMPACTSIZE(ADDR_IPV6_SIZE);
+                SerializeV1Stream(s);
+                return;
+            }
+
+            s << static_cast<uint8_t>(GetBIP155Network());
+            s << m_addr;
+        }
+
         /**
          * Unserialize from a pre-ADDRv2/BIP155 format from an array.
          */
@@ -272,6 +343,65 @@ class CNetAddr
 
             UnserializeV1Array(serialized);
         }
+
+        /**
+         * Unserialize from a ADDRv2 / BIP155 format.
+         */
+        template <typename Stream>
+        void UnserializeV2Stream(Stream& s)
+        {
+            uint8_t bip155_net;
+            s >> bip155_net;
+
+            size_t address_size;
+            s >> COMPACTSIZE(address_size);
+
+            if (address_size > MAX_ADDRV2_SIZE) {
+                throw std::ios_base::failure(strprintf(
+                    "Address too long: %u > %u", address_size, MAX_ADDRV2_SIZE));
+            }
+
+            scopeId = 0;
+
+            if (SetNetFromBIP155Network(bip155_net, address_size)) {
+                m_addr.resize(address_size);
+                s >> MakeSpan(m_addr);
+
+                if (m_net != NET_IPV6) {
+                    return;
+                }
+
+                // Do some special checks on IPv6 addresses.
+
+                // Recognize NET_INTERNAL embedded in IPv6, such addresses are not
+                // gossiped but could be coming from addrman, when unserializing from
+                // disk.
+                if (HasPrefix(m_addr, INTERNAL_IN_IPV6_PREFIX)) {
+                    m_net = NET_INTERNAL;
+                    memmove(m_addr.data(), m_addr.data() + INTERNAL_IN_IPV6_PREFIX.size(),
+                            ADDR_INTERNAL_SIZE);
+                    m_addr.resize(ADDR_INTERNAL_SIZE);
+                    return;
+                }
+
+                if (!HasPrefix(m_addr, IPV4_IN_IPV6_PREFIX) &&
+                    !HasPrefix(m_addr, TORV2_IN_IPV6_PREFIX)) {
+                    return;
+                }
+
+                // IPv4 and TORv2 are not supposed to be embedded in IPv6 (like in V1
+                // encoding). Unserialize as !IsValid(), thus ignoring them.
+            } else {
+                // If we receive an unknown BIP155 network id (from the future?) then
+                // ignore the address - unserialize as !IsValid().
+                s.ignore(address_size);
+            }
+
+            // Mimic a default-constructed CNetAddr object which is !IsValid() and thus
+            // will not be gossiped, but continue reading next addresses from the stream.
+            m_net = NET_IPV6;
+            m_addr.assign(ADDR_IPV6_SIZE, 0x0);
+        }
 };
 
 class CSubNet

src/primitives/transaction.h

@@ -14,6 +14,12 @@
 
 #include <tuple>
 
+/**
+ * A flag that is ORed into the protocol version to designate that a transaction
+ * should be (un)serialized without witness data.
+ * Make sure that this does not collide with any of the values in `version.h`
+ * or with `ADDRV2_FORMAT`.
+ */
 static const int SERIALIZE_TRANSACTION_NO_WITNESS = 0x40000000;
 
 /** An outpoint - a combination of a transaction hash and an index n into its vout */