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

Co-authored-by: Carl Dong <contact@carldong.me>

5 files changed, 389 insertions, 3 deletions

diff --git a/src/netaddress.cpp b/src/netaddress.cpp
index 86fca451f0..e60c2f1558 100644
--- a/src/netaddress.cpp
+++ b/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).
diff --git a/src/netaddress.h b/src/netaddress.h
index d00f5a6f55..08857ae729 100644
--- a/src/netaddress.h
+++ b/src/netaddress.h
@@ -13,13 +13,25 @@
 #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`
  * belongs to both `NET_UNROUTABLE` and `NET_IPV4`.
@@ -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);
+            }
         }
 
         /**
@@ -186,18 +202,54 @@ 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.
          */
@@ -251,6 +303,25 @@ class CNetAddr
         }
 
         /**
+         * 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.
          */
         void UnserializeV1Array(uint8_t (&arr)[V1_SERIALIZATION_SIZE])
@@ -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
diff --git a/src/primitives/transaction.h b/src/primitives/transaction.h
index 544bab6d9b..77cb1781a4 100644
--- a/src/primitives/transaction.h
+++ b/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 */
diff --git a/src/test/net_tests.cpp b/src/test/net_tests.cpp
index 85ebc89673..0bf3aa3e1f 100644
--- a/src/test/net_tests.cpp
+++ b/src/test/net_tests.cpp
@@ -10,6 +10,7 @@
 #include <net.h>
 #include <netbase.h>
 #include <serialize.h>
+#include <span.h>
 #include <streams.h>
 #include <test/util/setup_common.h>
 #include <util/memory.h>
@@ -20,6 +21,7 @@
 
 #include <boost/test/unit_test.hpp>
 
+#include <ios>
 #include <memory>
 #include <string>
 
@@ -261,17 +263,207 @@ BOOST_AUTO_TEST_CASE(cnetaddr_basic)
     BOOST_CHECK_EQUAL(addr.ToString(), "esffpvrt3wpeaygy.internal");
 }
 
-BOOST_AUTO_TEST_CASE(cnetaddr_serialize)
+BOOST_AUTO_TEST_CASE(cnetaddr_serialize_v1)
 {
     CNetAddr addr;
     CDataStream s(SER_NETWORK, PROTOCOL_VERSION);
 
+    s << addr;
+    BOOST_CHECK_EQUAL(HexStr(s), "00000000000000000000000000000000");
+    s.clear();
+
+    BOOST_REQUIRE(LookupHost("1.2.3.4", addr, false));
+    s << addr;
+    BOOST_CHECK_EQUAL(HexStr(s), "00000000000000000000ffff01020304");
+    s.clear();
+
+    BOOST_REQUIRE(LookupHost("1a1b:2a2b:3a3b:4a4b:5a5b:6a6b:7a7b:8a8b", addr, false));
+    s << addr;
+    BOOST_CHECK_EQUAL(HexStr(s), "1a1b2a2b3a3b4a4b5a5b6a6b7a7b8a8b");
+    s.clear();
+
+    BOOST_REQUIRE(addr.SetSpecial("6hzph5hv6337r6p2.onion"));
+    s << addr;
+    BOOST_CHECK_EQUAL(HexStr(s), "fd87d87eeb43f1f2f3f4f5f6f7f8f9fa");
+    s.clear();
+
     addr.SetInternal("a");
     s << addr;
     BOOST_CHECK_EQUAL(HexStr(s), "fd6b88c08724ca978112ca1bbdcafac2");
     s.clear();
 }
 
+BOOST_AUTO_TEST_CASE(cnetaddr_serialize_v2)
+{
+    CNetAddr addr;
+    CDataStream s(SER_NETWORK, PROTOCOL_VERSION);
+    // Add ADDRV2_FORMAT to the version so that the CNetAddr
+    // serialize method produces an address in v2 format.
+    s.SetVersion(s.GetVersion() | ADDRV2_FORMAT);
+
+    s << addr;
+    BOOST_CHECK_EQUAL(HexStr(s), "021000000000000000000000000000000000");
+    s.clear();
+
+    BOOST_REQUIRE(LookupHost("1.2.3.4", addr, false));
+    s << addr;
+    BOOST_CHECK_EQUAL(HexStr(s), "010401020304");
+    s.clear();
+
+    BOOST_REQUIRE(LookupHost("1a1b:2a2b:3a3b:4a4b:5a5b:6a6b:7a7b:8a8b", addr, false));
+    s << addr;
+    BOOST_CHECK_EQUAL(HexStr(s), "02101a1b2a2b3a3b4a4b5a5b6a6b7a7b8a8b");
+    s.clear();
+
+    BOOST_REQUIRE(addr.SetSpecial("6hzph5hv6337r6p2.onion"));
+    s << addr;
+    BOOST_CHECK_EQUAL(HexStr(s), "030af1f2f3f4f5f6f7f8f9fa");
+    s.clear();
+
+    BOOST_REQUIRE(addr.SetInternal("a"));
+    s << addr;
+    BOOST_CHECK_EQUAL(HexStr(s), "0210fd6b88c08724ca978112ca1bbdcafac2");
+    s.clear();
+}
+
+BOOST_AUTO_TEST_CASE(cnetaddr_unserialize_v2)
+{
+    CNetAddr addr;
+    CDataStream s(SER_NETWORK, PROTOCOL_VERSION);
+    // Add ADDRV2_FORMAT to the version so that the CNetAddr
+    // unserialize method expects an address in v2 format.
+    s.SetVersion(s.GetVersion() | ADDRV2_FORMAT);
+
+    // Valid IPv4.
+    s << MakeSpan(ParseHex("01"          // network type (IPv4)
+                           "04"          // address length
+                           "01020304")); // address
+    s >> addr;
+    BOOST_CHECK(addr.IsValid());
+    BOOST_CHECK(addr.IsIPv4());
+    BOOST_CHECK_EQUAL(addr.ToString(), "1.2.3.4");
+    BOOST_REQUIRE(s.empty());
+
+    // Invalid IPv4, valid length but address itself is shorter.
+    s << MakeSpan(ParseHex("01"      // network type (IPv4)
+                           "04"      // address length
+                           "0102")); // address
+    BOOST_CHECK_EXCEPTION(s >> addr, std::ios_base::failure, HasReason("end of data"));
+    BOOST_REQUIRE(!s.empty()); // The stream is not consumed on invalid input.
+    s.clear();
+
+    // Invalid IPv4, with bogus length.
+    s << MakeSpan(ParseHex("01"          // network type (IPv4)
+                           "05"          // address length
+                           "01020304")); // address
+    BOOST_CHECK_EXCEPTION(s >> addr, std::ios_base::failure,
+                          HasReason("BIP155 IPv4 address with length 5 (should be 4)"));
+    BOOST_REQUIRE(!s.empty()); // The stream is not consumed on invalid input.
+    s.clear();
+
+    // Invalid IPv4, with extreme length.
+    s << MakeSpan(ParseHex("01"          // network type (IPv4)
+                           "fd0102"      // address length (513 as CompactSize)
+                           "01020304")); // address
+    BOOST_CHECK_EXCEPTION(s >> addr, std::ios_base::failure,
+                          HasReason("Address too long: 513 > 512"));
+    BOOST_REQUIRE(!s.empty()); // The stream is not consumed on invalid input.
+    s.clear();
+
+    // Valid IPv6.
+    s << MakeSpan(ParseHex("02"                                  // network type (IPv6)
+                           "10"                                  // address length
+                           "0102030405060708090a0b0c0d0e0f10")); // address
+    s >> addr;
+    BOOST_CHECK(addr.IsValid());
+    BOOST_CHECK(addr.IsIPv6());
+    BOOST_CHECK_EQUAL(addr.ToString(), "102:304:506:708:90a:b0c:d0e:f10");
+    BOOST_REQUIRE(s.empty());
+
+    // Valid IPv6, contains embedded "internal".
+    s << MakeSpan(ParseHex(
+        "02"                                  // network type (IPv6)
+        "10"                                  // address length
+        "fd6b88c08724ca978112ca1bbdcafac2")); // address: 0xfd + sha256("bitcoin")[0:5] +
+                                              // sha256(name)[0:10]
+    s >> addr;
+    BOOST_CHECK(addr.IsInternal());
+    BOOST_CHECK_EQUAL(addr.ToString(), "zklycewkdo64v6wc.internal");
+    BOOST_REQUIRE(s.empty());
+
+    // Invalid IPv6, with bogus length.
+    s << MakeSpan(ParseHex("02"    // network type (IPv6)
+                           "04"    // address length
+                           "00")); // address
+    BOOST_CHECK_EXCEPTION(s >> addr, std::ios_base::failure,
+                          HasReason("BIP155 IPv6 address with length 4 (should be 16)"));
+    BOOST_REQUIRE(!s.empty()); // The stream is not consumed on invalid input.
+    s.clear();
+
+    // Invalid IPv6, contains embedded IPv4.
+    s << MakeSpan(ParseHex("02"                                  // network type (IPv6)
+                           "10"                                  // address length
+                           "00000000000000000000ffff01020304")); // address
+    s >> addr;
+    BOOST_CHECK(!addr.IsValid());
+    BOOST_REQUIRE(s.empty());
+
+    // Invalid IPv6, contains embedded TORv2.
+    s << MakeSpan(ParseHex("02"                                  // network type (IPv6)
+                           "10"                                  // address length
+                           "fd87d87eeb430102030405060708090a")); // address
+    s >> addr;
+    BOOST_CHECK(!addr.IsValid());
+    BOOST_REQUIRE(s.empty());
+
+    // Valid TORv2.
+    s << MakeSpan(ParseHex("03"                      // network type (TORv2)
+                           "0a"                      // address length
+                           "f1f2f3f4f5f6f7f8f9fa")); // address
+    s >> addr;
+    BOOST_CHECK(addr.IsValid());
+    BOOST_CHECK(addr.IsTor());
+    BOOST_CHECK_EQUAL(addr.ToString(), "6hzph5hv6337r6p2.onion");
+    BOOST_REQUIRE(s.empty());
+
+    // Invalid TORv2, with bogus length.
+    s << MakeSpan(ParseHex("03"    // network type (TORv2)
+                           "07"    // address length
+                           "00")); // address
+    BOOST_CHECK_EXCEPTION(s >> addr, std::ios_base::failure,
+                          HasReason("BIP155 TORv2 address with length 7 (should be 10)"));
+    BOOST_REQUIRE(!s.empty()); // The stream is not consumed on invalid input.
+    s.clear();
+
+    // Unknown, with extreme length.
+    s << MakeSpan(ParseHex("aa"             // network type (unknown)
+                           "fe00000002"     // address length (CompactSize's MAX_SIZE)
+                           "01020304050607" // address
+                           ));
+    BOOST_CHECK_EXCEPTION(s >> addr, std::ios_base::failure,
+                          HasReason("Address too long: 33554432 > 512"));
+    BOOST_REQUIRE(!s.empty()); // The stream is not consumed on invalid input.
+    s.clear();
+
+    // Unknown, with reasonable length.
+    s << MakeSpan(ParseHex("aa"       // network type (unknown)
+                           "04"       // address length
+                           "01020304" // address
+                           ));
+    s >> addr;
+    BOOST_CHECK(!addr.IsValid());
+    BOOST_REQUIRE(s.empty());
+
+    // Unknown, with zero length.
+    s << MakeSpan(ParseHex("aa" // network type (unknown)
+                           "00" // address length
+                           ""   // address
+                           ));
+    s >> addr;
+    BOOST_CHECK(!addr.IsValid());
+    BOOST_REQUIRE(s.empty());
+}
+
 // prior to PR #14728, this test triggers an undefined behavior
 BOOST_AUTO_TEST_CASE(ipv4_peer_with_ipv6_addrMe_test)
 {
diff --git a/src/version.h b/src/version.h
index b5f379e1b8..019c3a3ae7 100644
--- a/src/version.h
+++ b/src/version.h
@@ -38,4 +38,7 @@ 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;
 
+// Make sure that none of the values above collide with
+// `SERIALIZE_TRANSACTION_NO_WITNESS` or `ADDRV2_FORMAT`.
+
 #endif // BITCOIN_VERSION_H