rpc: Enhance metadata of the dumptxoutset output

The following data is added:
- A newly introduced utxo set magic
- A version number
- The network magic
- The block height

Author: fjahr

src/kernel/chainparams.cpp

@@ -542,3 +542,33 @@ std::unique_ptr<const CChainParams> CChainParams::TestNet()
 {
     return std::make_unique<const CTestNetParams>();
 }
+
+std::vector<int> CChainParams::GetAvailableSnapshotHeights() const
+{
+    std::vector<int> heights;
+    heights.reserve(m_assumeutxo_data.size());
+
+    for (const auto& data : m_assumeutxo_data) {
+        heights.emplace_back(data.height);
+    }
+    return heights;
+}
+
+std::optional<ChainType> GetNetworkForMagic(MessageStartChars& message)
+{
+    const auto mainnet_msg = CChainParams::Main()->MessageStart();
+    const auto testnet_msg = CChainParams::TestNet()->MessageStart();
+    const auto regtest_msg = CChainParams::RegTest({})->MessageStart();
+    const auto signet_msg = CChainParams::SigNet({})->MessageStart();
+
+    if (std::equal(message.begin(), message.end(), mainnet_msg.data())) {
+        return ChainType::MAIN;
+    } else if (std::equal(message.begin(), message.end(), testnet_msg.data())) {
+        return ChainType::TESTNET;
+    } else if (std::equal(message.begin(), message.end(), regtest_msg.data())) {
+        return ChainType::REGTEST;
+    } else if (std::equal(message.begin(), message.end(), signet_msg.data())) {
+        return ChainType::SIGNET;
+    }
+    return std::nullopt;
+}

src/kernel/chainparams.h

@@ -93,6 +93,7 @@ class CChainParams
     const Consensus::Params& GetConsensus() const { return consensus; }
     const MessageStartChars& MessageStart() const { return pchMessageStart; }
     uint16_t GetDefaultPort() const { return nDefaultPort; }
+    std::vector<int> GetAvailableSnapshotHeights() const;
 
     const CBlock& GenesisBlock() const { return genesis; }
     /** Default value for -checkmempool and -checkblockindex argument */
@@ -183,4 +184,6 @@ class CChainParams
     ChainTxData chainTxData;
 };
 
+std::optional<ChainType> GetNetworkForMagic(MessageStartChars& pchMessageStart);
+
 #endif // BITCOIN_KERNEL_CHAINPARAMS_H

src/node/utxo_snapshot.h

@@ -6,27 +6,37 @@
 #ifndef BITCOIN_NODE_UTXO_SNAPSHOT_H
 #define BITCOIN_NODE_UTXO_SNAPSHOT_H
 
+#include <chainparams.h>
+#include <kernel/chainparams.h>
 #include <kernel/cs_main.h>
 #include <serialize.h>
 #include <sync.h>
 #include <uint256.h>
+#include <util/chaintype.h>
 #include <util/fs.h>
 
 #include <cstdint>
 #include <optional>
 #include <string_view>
 
+// UTXO set snapshot magic bytes
+static constexpr std::array<uint8_t, 5> SNAPSHOT_MAGIC_BYTES = {'u', 't', 'x', 'o', 0xff};
+
 class Chainstate;
 
 namespace node {
 //! Metadata describing a serialized version of a UTXO set from which an
 //! assumeutxo Chainstate can be constructed.
 class SnapshotMetadata
 {
+    const uint16_t m_version{1};
+    const std::set<uint16_t> m_supported_versions{1};
 public:
     //! The hash of the block that reflects the tip of the chain for the
     //! UTXO set contained in this snapshot.
     uint256 m_base_blockhash;
+    uint32_t m_base_blockheight;
+
 
     //! The number of coins in the UTXO set contained in this snapshot. Used
     //! during snapshot load to estimate progress of UTXO set reconstruction.
@@ -35,11 +45,55 @@ class SnapshotMetadata
     SnapshotMetadata() { }
     SnapshotMetadata(
         const uint256& base_blockhash,
+        const int base_blockheight,
         uint64_t coins_count) :
             m_base_blockhash(base_blockhash),
+            m_base_blockheight(base_blockheight),
             m_coins_count(coins_count) { }
 
-    SERIALIZE_METHODS(SnapshotMetadata, obj) { READWRITE(obj.m_base_blockhash, obj.m_coins_count); }
+    template <typename Stream>
+    inline void Serialize(Stream& s) const {
+        s << SNAPSHOT_MAGIC_BYTES;
+        s << m_version;
+        s << Params().MessageStart();
+        s << m_base_blockheight;
+        s << m_base_blockhash;
+        s << m_coins_count;
+    }
+
+    template <typename Stream>
+    inline void Unserialize(Stream& s) {
+        // Read the snapshot magic bytes
+        std::array<uint8_t, SNAPSHOT_MAGIC_BYTES.size()> snapshot_magic;
+        s >> snapshot_magic;
+        if (snapshot_magic != SNAPSHOT_MAGIC_BYTES) {
+            throw std::ios_base::failure("Invalid UTXO set snapshot magic bytes. Please check if this is indeed a snapshot file or if you are using an outdated snapshot format.");
+        }
+
+        // Read the version
+        uint16_t version;
+        s >> version;
+        if (m_supported_versions.find(version) == m_supported_versions.end()) {
+            throw std::ios_base::failure(strprintf("Version of snapshot %s does not match any of the supported versions.", version));
+        }
+
+        // Read the network magic (pchMessageStart)
+        MessageStartChars message;
+        s >> message;
+        if (!std::equal(message.begin(), message.end(), Params().MessageStart().data())) {
+            auto metadata_network = GetNetworkForMagic(message);
+            if (metadata_network) {
+                std::string network_string{ChainTypeToString(metadata_network.value())};
+                throw std::ios_base::failure(strprintf("The network of the snapshot (%s) does not match the network of this node (%s).", network_string, Params().GetChainTypeString()));
+            } else {
+                throw std::ios_base::failure("This snapshot has been created for an unrecognized network. This could be a custom signet, a new testnet or possibly caused by data corruption.");
+            }
+        }
+
+        s >> m_base_blockheight;
+        s >> m_base_blockhash;
+        s >> m_coins_count;
+    }
 };
 
 //! The file in the snapshot chainstate dir which stores the base blockhash. This is

src/rpc/blockchain.cpp

@@ -2691,7 +2691,7 @@ UniValue CreateUTXOSnapshot(
         tip->nHeight, tip->GetBlockHash().ToString(),
         fs::PathToString(path), fs::PathToString(temppath)));
 
-    SnapshotMetadata metadata{tip->GetBlockHash(), maybe_stats->coins_count};
+    SnapshotMetadata metadata{tip->GetBlockHash(), tip->nHeight, maybe_stats->coins_count};
 
     afile << metadata;
 
@@ -2804,12 +2804,22 @@ static RPCHelpMan loadtxoutset()
     }
 
     SnapshotMetadata metadata;
-    afile >> metadata;
+    try {
+        afile >> metadata;
+    } catch (const std::ios_base::failure& e) {
+        throw JSONRPCError(RPC_DESERIALIZATION_ERROR, strprintf("Unable to parse metadata: %s", e.what()));
+    }
 
     uint256 base_blockhash = metadata.m_base_blockhash;
+    int base_blockheight = metadata.m_base_blockheight;
     if (!chainman.GetParams().AssumeutxoForBlockhash(base_blockhash).has_value()) {
+        auto available_heights = chainman.GetParams().GetAvailableSnapshotHeights();
+        std::string heights_formatted = Join(available_heights, ", ", [&](const auto& i) { return ToString(i); });
         throw JSONRPCError(RPC_INTERNAL_ERROR, strprintf("Unable to load UTXO snapshot, "
-            "assumeutxo block hash in snapshot metadata not recognized (%s)", base_blockhash.ToString()));
+            "assumeutxo block hash in snapshot metadata not recognized (hash: %s, height: %s). The following snapshot heights are available: %s.",
+            base_blockhash.ToString(),
+            base_blockheight,
+            heights_formatted));
     }
     CBlockIndex* snapshot_start_block = WITH_LOCK(::cs_main,
             return chainman.m_blockman.LookupBlockIndex(base_blockhash));

test/functional/feature_assumeutxo.py

@@ -75,26 +75,57 @@ def expected_error(log_msg="", rpc_details=""):
             with self.nodes[1].assert_debug_log([log_msg]):
                 assert_raises_rpc_error(-32603, f"Unable to load UTXO snapshot{rpc_details}", self.nodes[1].loadtxoutset, bad_snapshot_path)
 
+        self.log.info("  - snapshot file with invalid file magic")
+        parsing_error_code = -22
+        bad_magic = 0xf00f00f000
+        with open(bad_snapshot_path, 'wb') as f:
+            f.write(bad_magic.to_bytes(5, "big") + valid_snapshot_contents[5:])
+        assert_raises_rpc_error(parsing_error_code, "Unable to parse metadata: Invalid UTXO set snapshot magic bytes. Please check if this is indeed a snapshot file or if you are using an outdated snapshot format.", self.nodes[1].loadtxoutset, bad_snapshot_path)
+
+        self.log.info("  - snapshot file with unsupported version")
+        for version in [0, 2]:
+            with open(bad_snapshot_path, 'wb') as f:
+                f.write(valid_snapshot_contents[:5] + version.to_bytes(2, "little") + valid_snapshot_contents[7:])
+            assert_raises_rpc_error(parsing_error_code, f"Unable to parse metadata: Version of snapshot {version} does not match any of the supported versions.", self.nodes[1].loadtxoutset, bad_snapshot_path)
+
+        self.log.info("  - snapshot file with mismatching network magic")
+        invalid_magics = [
+            # magic, name, real
+            [0xf9beb4d9, "main", True],
+            [0x0b110907, "test", True],
+            [0x0a03cf40, "signet", True],
+            [0x00000000, "", False],
+            [0xffffffff, "", False],
+        ]
+        for [magic, name, real] in invalid_magics:
+            with open(bad_snapshot_path, 'wb') as f:
+                f.write(valid_snapshot_contents[:7] + magic.to_bytes(4, 'big') + valid_snapshot_contents[11:])
+            if real:
+                assert_raises_rpc_error(parsing_error_code, f"Unable to parse metadata: The network of the snapshot ({name}) does not match the network of this node (regtest).", self.nodes[1].loadtxoutset, bad_snapshot_path)
+            else:
+                assert_raises_rpc_error(parsing_error_code, "Unable to parse metadata: This snapshot has been created for an unrecognized network. This could be a custom signet, a new testnet or possibly caused by data corruption.", self.nodes[1].loadtxoutset, bad_snapshot_path)
+
         self.log.info("  - snapshot file referring to a block that is not in the assumeutxo parameters")
         prev_block_hash = self.nodes[0].getblockhash(SNAPSHOT_BASE_HEIGHT - 1)
         bogus_block_hash = "0" * 64  # Represents any unknown block hash
+        # The height is not used for anything critical currently, so we just
+        # confirm the manipulation in the error message
+        bogus_height = 1337
         for bad_block_hash in [bogus_block_hash, prev_block_hash]:
             with open(bad_snapshot_path, 'wb') as f:
-                # block hash of the snapshot base is stored right at the start (first 32 bytes)
-                f.write(bytes.fromhex(bad_block_hash)[::-1] + valid_snapshot_contents[32:])
-            error_details = f", assumeutxo block hash in snapshot metadata not recognized ({bad_block_hash})"
+                f.write(valid_snapshot_contents[:11] + bogus_height.to_bytes(4, "little") + bytes.fromhex(bad_block_hash)[::-1] + valid_snapshot_contents[47:])
+            error_details = f", assumeutxo block hash in snapshot metadata not recognized (hash: {bad_block_hash}, height: {bogus_height}). The following snapshot heights are available: 110, 299."
             expected_error(rpc_details=error_details)
 
         self.log.info("  - snapshot file with wrong number of coins")
-        valid_num_coins = int.from_bytes(valid_snapshot_contents[32:32 + 8], "little")
+        valid_num_coins = int.from_bytes(valid_snapshot_contents[47:47 + 8], "little")
         for off in [-1, +1]:
             with open(bad_snapshot_path, 'wb') as f:
-                f.write(valid_snapshot_contents[:32])
+                f.write(valid_snapshot_contents[:47])
                 f.write((valid_num_coins + off).to_bytes(8, "little"))
-                f.write(valid_snapshot_contents[32 + 8:])
+                f.write(valid_snapshot_contents[47 + 8:])
             expected_error(log_msg=f"bad snapshot - coins left over after deserializing 298 coins" if off == -1 else f"bad snapshot format or truncated snapshot after deserializing 299 coins")
 
-
         self.log.info("  - snapshot file with alternated but parsable UTXO data results in different hash")
         cases = [
             # (content, offset, wrong_hash, custom_message)
@@ -109,9 +140,10 @@ def expected_error(log_msg="", rpc_details=""):
 
         for content, offset, wrong_hash, custom_message in cases:
             with open(bad_snapshot_path, "wb") as f:
-                f.write(valid_snapshot_contents[:(32 + 8 + offset)])
+                # Prior to offset: Snapshot magic, snapshot version, network magic, height, hash, coins count
+                f.write(valid_snapshot_contents[:(5 + 2 + 4 + 4 + 32 + 8 + offset)])
                 f.write(content)
-                f.write(valid_snapshot_contents[(32 + 8 + offset + len(content)):])
+                f.write(valid_snapshot_contents[(5 + 2 + 4 + 4 + 32 + 8 + offset + len(content)):])
 
             log_msg = custom_message if custom_message is not None else f"[snapshot] bad snapshot content hash: expected a4bf3407ccb2cc0145c49ebba8fa91199f8a3903daf0883875941497d2493c27, got {wrong_hash}"
             expected_error(log_msg=log_msg)

test/functional/rpc_dumptxoutset.py

@@ -43,7 +43,7 @@ def run_test(self):
         # UTXO snapshot hash should be deterministic based on mocked time.
         assert_equal(
             sha256sum_file(str(expected_path)).hex(),
-            '3263fc0311ea46415b85513a59ad8fe67806b3cdce66147175ecb9da768d4a99')
+            '2f775f82811150d310527b5ff773f81fb0fb517e941c543c1f7c4d38fd2717b3')
 
         assert_equal(
             out['txoutset_hash'], 'a0b7baa3bf5ccbd3279728f230d7ca0c44a76e9923fca8f32dbfd08d65ea496a')