eth/filters, cmd: add config of eth_getLogs address limit (#32327)

Add cli configurable limit for the number of addresses allowed in
eth_getLogs filter criteria:
https://github.com/ethereum/go-ethereum/issues/32264
 
 Key changes:
- Added --rpc.getlogmaxaddrs CLI flag (default: 1000) to configure the
maximum number of addresses
- Updated ethconfig.Config with FilterMaxAddresses field for
configuration management
- Modified filter system to use the configurable limit instead of the
hardcoded maxAddresses constant
- Enhanced test coverage with new test cases for address limit
validation
- Removed hardcoded validation from JSON unmarshaling, moving it to
runtime validation

Please notice that I remove the check at FilterCriteria UnmarshalJSON
because the runtime config can not pass into this validation.

Please help review this change!

---------

Co-authored-by: zsfelfoldi <zsfelfoldi@gmail.com>
Co-authored-by: rjl493456442 <garyrong0905@gmail.com>

cmd/geth/main.go

@@ -182,6 +182,7 @@ var (
 		utils.RPCGlobalGasCapFlag,
 		utils.RPCGlobalEVMTimeoutFlag,
 		utils.RPCGlobalTxFeeCapFlag,
+		utils.RPCGlobalLogQueryLimit,
 		utils.AllowUnprotectedTxs,
 		utils.BatchRequestLimit,
 		utils.BatchResponseMaxSize,

cmd/utils/flags.go

@@ -600,6 +600,12 @@ var (
 		Value:    ethconfig.Defaults.RPCTxFeeCap,
 		Category: flags.APICategory,
 	}
+	RPCGlobalLogQueryLimit = &cli.IntFlag{
+		Name:     "rpc.logquerylimit",
+		Usage:    "Maximum number of alternative addresses or topics allowed per search position in eth_getLogs filter criteria (0 = no cap)",
+		Value:    ethconfig.Defaults.LogQueryLimit,
+		Category: flags.APICategory,
+	}
 	// Authenticated RPC HTTP settings
 	AuthListenFlag = &cli.StringFlag{
 		Name:     "authrpc.addr",
@@ -1699,6 +1705,9 @@ func SetEthConfig(ctx *cli.Context, stack *node.Node, cfg *ethconfig.Config) {
 	if ctx.IsSet(CacheLogSizeFlag.Name) {
 		cfg.FilterLogCacheSize = ctx.Int(CacheLogSizeFlag.Name)
 	}
+	if ctx.IsSet(RPCGlobalLogQueryLimit.Name) {
+		cfg.LogQueryLimit = ctx.Int(RPCGlobalLogQueryLimit.Name)
+	}
 	if !ctx.Bool(SnapshotFlag.Name) || cfg.SnapshotCache == 0 {
 		// If snap-sync is requested, this flag is also required
 		if cfg.SyncMode == ethconfig.SnapSync {
@@ -2017,7 +2026,8 @@ func RegisterGraphQLService(stack *node.Node, backend ethapi.Backend, filterSyst
 // RegisterFilterAPI adds the eth log filtering RPC API to the node.
 func RegisterFilterAPI(stack *node.Node, backend ethapi.Backend, ethcfg *ethconfig.Config) *filters.FilterSystem {
 	filterSystem := filters.NewFilterSystem(backend, filters.Config{
-		LogCacheSize: ethcfg.FilterLogCacheSize,
+		LogCacheSize:  ethcfg.FilterLogCacheSize,
+		LogQueryLimit: ethcfg.LogQueryLimit,
 	})
 	stack.RegisterAPIs([]rpc.API{{
 		Namespace: "eth",

eth/ethconfig/config.go

@@ -62,6 +62,7 @@ var Defaults = Config{
 	TrieTimeout:        60 * time.Minute,
 	SnapshotCache:      102,
 	FilterLogCacheSize: 32,
+	LogQueryLimit:      1000,
 	Miner:              miner.DefaultConfig,
 	TxPool:             legacypool.DefaultConfig,
 	BlobPool:           blobpool.DefaultConfig,
@@ -131,6 +132,10 @@ type Config struct {
 	// This is the number of blocks for which logs will be cached in the filter system.
 	FilterLogCacheSize int
 
+	// This is the maximum number of addresses or topics allowed in filter criteria
+	// for eth_getLogs.
+	LogQueryLimit int
+
 	// Mining options
 	Miner miner.Config
 

eth/ethconfig/gen_config.go

@@ -44,6 +44,7 @@ func (c Config) MarshalTOML() (interface{}, error) {
 		SnapshotCache           int
 		Preimages               bool
 		FilterLogCacheSize      int
+		LogQueryLimit           int
 		Miner                   miner.Config
 		TxPool                  legacypool.Config
 		BlobPool                blobpool.Config
@@ -88,6 +89,7 @@ func (c Config) MarshalTOML() (interface{}, error) {
 	enc.SnapshotCache = c.SnapshotCache
 	enc.Preimages = c.Preimages
 	enc.FilterLogCacheSize = c.FilterLogCacheSize
+	enc.LogQueryLimit = c.LogQueryLimit
 	enc.Miner = c.Miner
 	enc.TxPool = c.TxPool
 	enc.BlobPool = c.BlobPool
@@ -136,6 +138,7 @@ func (c *Config) UnmarshalTOML(unmarshal func(interface{}) error) error {
 		SnapshotCache           *int
 		Preimages               *bool
 		FilterLogCacheSize      *int
+		LogQueryLimit           *int
 		Miner                   *miner.Config
 		TxPool                  *legacypool.Config
 		BlobPool                *blobpool.Config
@@ -237,6 +240,9 @@ func (c *Config) UnmarshalTOML(unmarshal func(interface{}) error) error {
 	if dec.FilterLogCacheSize != nil {
 		c.FilterLogCacheSize = *dec.FilterLogCacheSize
 	}
+	if dec.LogQueryLimit != nil {
+		c.LogQueryLimit = *dec.LogQueryLimit
+	}
 	if dec.Miner != nil {
 		c.Miner = *dec.Miner
 	}

eth/filters/api.go

@@ -42,12 +42,10 @@ var (
 	errBlockHashWithRange     = errors.New("can't specify fromBlock/toBlock with blockHash")
 	errPendingLogsUnsupported = errors.New("pending logs are not supported")
 	errExceedMaxTopics        = errors.New("exceed max topics")
-	errExceedMaxAddresses     = errors.New("exceed max addresses")
+	errExceedLogQueryLimit    = errors.New("exceed max addresses or topics per search position")
 )
 
 const (
-	// The maximum number of addresses allowed in a filter criteria
-	maxAddresses = 1000
 	// The maximum number of topic criteria allowed, vm.LOG4 - vm.LOG0
 	maxTopics = 4
 	// The maximum number of allowed topics within a topic criteria
@@ -70,20 +68,22 @@ type filter struct {
 // FilterAPI offers support to create and manage filters. This will allow external clients to retrieve various
 // information related to the Ethereum protocol such as blocks, transactions and logs.
 type FilterAPI struct {
-	sys       *FilterSystem
-	events    *EventSystem
-	filtersMu sync.Mutex
-	filters   map[rpc.ID]*filter
-	timeout   time.Duration
+	sys           *FilterSystem
+	events        *EventSystem
+	filtersMu     sync.Mutex
+	filters       map[rpc.ID]*filter
+	timeout       time.Duration
+	logQueryLimit int
 }
 
 // NewFilterAPI returns a new FilterAPI instance.
 func NewFilterAPI(system *FilterSystem) *FilterAPI {
 	api := &FilterAPI{
-		sys:     system,
-		events:  NewEventSystem(system),
-		filters: make(map[rpc.ID]*filter),
-		timeout: system.cfg.Timeout,
+		sys:           system,
+		events:        NewEventSystem(system),
+		filters:       make(map[rpc.ID]*filter),
+		timeout:       system.cfg.Timeout,
+		logQueryLimit: system.cfg.LogQueryLimit,
 	}
 	go api.timeoutLoop(system.cfg.Timeout)
 
@@ -347,8 +347,15 @@ func (api *FilterAPI) GetLogs(ctx context.Context, crit FilterCriteria) ([]*type
 	if len(crit.Topics) > maxTopics {
 		return nil, errExceedMaxTopics
 	}
-	if len(crit.Addresses) > maxAddresses {
-		return nil, errExceedMaxAddresses
+	if api.logQueryLimit != 0 {
+		if len(crit.Addresses) > api.logQueryLimit {
+			return nil, errExceedLogQueryLimit
+		}
+		for _, topics := range crit.Topics {
+			if len(topics) > api.logQueryLimit {
+				return nil, errExceedLogQueryLimit
+			}
+		}
 	}
 
 	var filter *Filter
@@ -545,9 +552,6 @@ func (args *FilterCriteria) UnmarshalJSON(data []byte) error {
 		// raw.Address can contain a single address or an array of addresses
 		switch rawAddr := raw.Addresses.(type) {
 		case []interface{}:
-			if len(rawAddr) > maxAddresses {
-				return errExceedMaxAddresses
-			}
 			for i, addr := range rawAddr {
 				if strAddr, ok := addr.(string); ok {
 					addr, err := decodeAddress(strAddr)

eth/filters/api_test.go

@@ -19,7 +19,6 @@ package filters
 import (
 	"encoding/json"
 	"fmt"
-	"strings"
 	"testing"
 
 	"github.com/ethereum/go-ethereum/common"
@@ -183,15 +182,4 @@ func TestUnmarshalJSONNewFilterArgs(t *testing.T) {
 	if len(test7.Topics[2]) != 0 {
 		t.Fatalf("expected 0 topics, got %d topics", len(test7.Topics[2]))
 	}
-
-	// multiple address exceeding max
-	var test8 FilterCriteria
-	addresses := make([]string, maxAddresses+1)
-	for i := 0; i < maxAddresses+1; i++ {
-		addresses[i] = fmt.Sprintf(`"%s"`, common.HexToAddress(fmt.Sprintf("0x%x", i)).Hex())
-	}
-	vector = fmt.Sprintf(`{"address": [%s]}`, strings.Join(addresses, ", "))
-	if err := json.Unmarshal([]byte(vector), &test8); err != errExceedMaxAddresses {
-		t.Fatal("expected errExceedMaxAddresses, got", err)
-	}
 }

eth/filters/filter_system.go

@@ -41,8 +41,9 @@ import (
 
 // Config represents the configuration of the filter system.
 type Config struct {
-	LogCacheSize int           // maximum number of cached blocks (default: 32)
-	Timeout      time.Duration // how long filters stay active (default: 5min)
+	LogCacheSize  int           // maximum number of cached blocks (default: 32)
+	Timeout       time.Duration // how long filters stay active (default: 5min)
+	LogQueryLimit int           // maximum number of addresses allowed in filter criteria (default: 1000)
 }
 
 func (cfg Config) withDefaults() Config {
@@ -291,8 +292,15 @@ func (es *EventSystem) SubscribeLogs(crit ethereum.FilterQuery, logs chan []*typ
 	if len(crit.Topics) > maxTopics {
 		return nil, errExceedMaxTopics
 	}
-	if len(crit.Addresses) > maxAddresses {
-		return nil, errExceedMaxAddresses
+	if es.sys.cfg.LogQueryLimit != 0 {
+		if len(crit.Addresses) > es.sys.cfg.LogQueryLimit {
+			return nil, errExceedLogQueryLimit
+		}
+		for _, topics := range crit.Topics {
+			if len(topics) > es.sys.cfg.LogQueryLimit {
+				return nil, errExceedLogQueryLimit
+			}
+		}
 	}
 	var from, to rpc.BlockNumber
 	if crit.FromBlock == nil {

eth/filters/filter_system_test.go

@@ -36,6 +36,7 @@ import (
 	"github.com/ethereum/go-ethereum/internal/ethapi"
 	"github.com/ethereum/go-ethereum/params"
 	"github.com/ethereum/go-ethereum/rpc"
+	"github.com/ethereum/go-ethereum/triedb"
 )
 
 type testBackend struct {
@@ -424,7 +425,7 @@ func TestInvalidLogFilterCreation(t *testing.T) {
 
 	var (
 		db     = rawdb.NewMemoryDatabase()
-		_, sys = newTestFilterSystem(db, Config{})
+		_, sys = newTestFilterSystem(db, Config{LogQueryLimit: 1000})
 		api    = NewFilterAPI(sys)
 	)
 
@@ -435,7 +436,7 @@ func TestInvalidLogFilterCreation(t *testing.T) {
 		1: {FromBlock: big.NewInt(rpc.PendingBlockNumber.Int64()), ToBlock: big.NewInt(100)},
 		2: {FromBlock: big.NewInt(rpc.LatestBlockNumber.Int64()), ToBlock: big.NewInt(100)},
 		3: {Topics: [][]common.Hash{{}, {}, {}, {}, {}}},
-		4: {Addresses: make([]common.Address, maxAddresses+1)},
+		4: {Addresses: make([]common.Address, api.logQueryLimit+1)},
 	}
 
 	for i, test := range testCases {
@@ -455,7 +456,7 @@ func TestInvalidGetLogsRequest(t *testing.T) {
 			BaseFee: big.NewInt(params.InitialBaseFee),
 		}
 		db, blocks, _    = core.GenerateChainWithGenesis(genesis, ethash.NewFaker(), 10, func(i int, gen *core.BlockGen) {})
-		_, sys           = newTestFilterSystem(db, Config{})
+		_, sys           = newTestFilterSystem(db, Config{LogQueryLimit: 10})
 		api              = NewFilterAPI(sys)
 		blockHash        = blocks[0].Hash()
 		unknownBlockHash = common.HexToHash("0x1111111111111111111111111111111111111111111111111111111111111111")
@@ -500,8 +501,8 @@ func TestInvalidGetLogsRequest(t *testing.T) {
 			err: errExceedMaxTopics,
 		},
 		{
-			f:   FilterCriteria{BlockHash: &blockHash, Addresses: make([]common.Address, maxAddresses+1)},
-			err: errExceedMaxAddresses,
+			f:   FilterCriteria{BlockHash: &blockHash, Addresses: make([]common.Address, api.logQueryLimit+1)},
+			err: errExceedLogQueryLimit,
 		},
 	}
 
@@ -528,6 +529,92 @@ func TestInvalidGetRangeLogsRequest(t *testing.T) {
 	}
 }
 
+// TestExceedLogQueryLimit tests getLogs with too many addresses or topics
+func TestExceedLogQueryLimit(t *testing.T) {
+	t.Parallel()
+
+	// Test with custom config (LogQueryLimit = 5 for easier testing)
+	var (
+		db           = rawdb.NewMemoryDatabase()
+		backend, sys = newTestFilterSystem(db, Config{LogQueryLimit: 5})
+		api          = NewFilterAPI(sys)
+		gspec        = &core.Genesis{
+			Config:  params.TestChainConfig,
+			Alloc:   types.GenesisAlloc{},
+			BaseFee: big.NewInt(params.InitialBaseFee),
+		}
+	)
+
+	_, err := gspec.Commit(db, triedb.NewDatabase(db, nil))
+	if err != nil {
+		t.Fatal(err)
+	}
+	chain, _ := core.GenerateChain(gspec.Config, gspec.ToBlock(), ethash.NewFaker(), db, 1000, func(i int, gen *core.BlockGen) {})
+
+	options := core.DefaultConfig().WithStateScheme(rawdb.HashScheme)
+	options.TxLookupLimit = 0 // index all txs
+	bc, err := core.NewBlockChain(db, gspec, ethash.NewFaker(), options)
+	if err != nil {
+		t.Fatal(err)
+	}
+	_, err = bc.InsertChain(chain[:600])
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	backend.startFilterMaps(200, false, filtermaps.RangeTestParams)
+	defer backend.stopFilterMaps()
+
+	addresses := make([]common.Address, 6)
+	for i := range addresses {
+		addresses[i] = common.HexToAddress("0x1234567890123456789012345678901234567890")
+	}
+
+	topics := make([]common.Hash, 6)
+	for i := range topics {
+		topics[i] = common.HexToHash("0x123456789012345678901234567890123456789001234567890012345678901234")
+	}
+
+	// Test that 5 addresses do not result in error
+	// Add FromBlock and ToBlock to make it similar to other invalid tests
+	if _, err := api.GetLogs(context.Background(), FilterCriteria{
+		FromBlock: big.NewInt(0),
+		ToBlock:   big.NewInt(100),
+		Addresses: addresses[:5],
+	}); err != nil {
+		t.Errorf("Expected GetLogs with 5 addresses to return with no error, got: %v", err)
+	}
+
+	// Test that 6 addresses fails with correct error
+	if _, err := api.GetLogs(context.Background(), FilterCriteria{
+		FromBlock: big.NewInt(0),
+		ToBlock:   big.NewInt(100),
+		Addresses: addresses,
+	}); err != errExceedLogQueryLimit {
+		t.Errorf("Expected GetLogs with 6 addresses to return errExceedLogQueryLimit, got: %v", err)
+	}
+
+	// Test that 5 topics at one position do not result in error
+	if _, err := api.GetLogs(context.Background(), FilterCriteria{
+		FromBlock: big.NewInt(0),
+		ToBlock:   big.NewInt(100),
+		Addresses: addresses[:1],
+		Topics:    [][]common.Hash{topics[:5]},
+	}); err != nil {
+		t.Errorf("Expected GetLogs with 5 topics at one position to return with no error, got: %v", err)
+	}
+
+	// Test that 6 topics at one position fails with correct error
+	if _, err := api.GetLogs(context.Background(), FilterCriteria{
+		FromBlock: big.NewInt(0),
+		ToBlock:   big.NewInt(100),
+		Addresses: addresses[:1],
+		Topics:    [][]common.Hash{topics},
+	}); err != errExceedLogQueryLimit {
+		t.Errorf("Expected GetLogs with 6 topics at one position to return errExceedLogQueryLimit, got: %v", err)
+	}
+}
+
 // TestLogFilter tests whether log filters match the correct logs that are posted to the event feed.
 func TestLogFilter(t *testing.T) {
 	t.Parallel()