swarm/network, swarm/storage: validate chunk size (#17397)

* swarm/network, swarm/storage: validate default chunk size

* swarm/bmt, swarm/network, swarm/storage: update BMT hash initialisation

* swarm/bmt: move segmentCount to tests

* swarm/chunk: change chunk.DefaultSize to be untyped const

* swarm/storage: add size validator

* swarm/storage: add chunk size validation to localstore

* swarm/storage: move validation from localstore to validator

* swarm/storage: global chunk rules in MRU

11 files changed, 49 insertions, 57 deletions

diff --git a/swarm/storage/chunker.go b/swarm/storage/chunker.go
index b9b502273..6d805b8e2 100644
--- a/swarm/storage/chunker.go
+++ b/swarm/storage/chunker.go
@@ -25,6 +25,7 @@ import (
 	"time"
 
 	"github.com/ethereum/go-ethereum/metrics"
+	"github.com/ethereum/go-ethereum/swarm/chunk"
 	"github.com/ethereum/go-ethereum/swarm/log"
 	"github.com/ethereum/go-ethereum/swarm/spancontext"
 	opentracing "github.com/opentracing/opentracing-go"
@@ -69,10 +70,6 @@ var (
 	errOperationTimedOut    = errors.New("operation timed out")
 )
 
-const (
-	DefaultChunkSize int64 = 4096
-)
-
 type ChunkerParams struct {
 	chunkSize int64
 	hashSize  int64
@@ -136,7 +133,7 @@ type TreeChunker struct {
 func TreeJoin(ctx context.Context, addr Address, getter Getter, depth int) *LazyChunkReader {
 	jp := &JoinerParams{
 		ChunkerParams: ChunkerParams{
-			chunkSize: DefaultChunkSize,
+			chunkSize: chunk.DefaultSize,
 			hashSize:  int64(len(addr)),
 		},
 		addr:   addr,
@@ -156,7 +153,7 @@ func TreeSplit(ctx context.Context, data io.Reader, size int64, putter Putter) (
 	tsp := &TreeSplitterParams{
 		SplitterParams: SplitterParams{
 			ChunkerParams: ChunkerParams{
-				chunkSize: DefaultChunkSize,
+				chunkSize: chunk.DefaultSize,
 				hashSize:  putter.RefSize(),
 			},
 			reader: data,
diff --git a/swarm/storage/hasherstore.go b/swarm/storage/hasherstore.go
index 139c0ee03..bc23077c1 100644
--- a/swarm/storage/hasherstore.go
+++ b/swarm/storage/hasherstore.go
@@ -22,6 +22,7 @@ import (
 	"sync"
 
 	"github.com/ethereum/go-ethereum/crypto/sha3"
+	"github.com/ethereum/go-ethereum/swarm/chunk"
 	"github.com/ethereum/go-ethereum/swarm/storage/encryption"
 )
 
@@ -57,7 +58,7 @@ func NewHasherStore(chunkStore ChunkStore, hashFunc SwarmHasher, toEncrypt bool)
 	refSize := int64(hashSize)
 	if toEncrypt {
 		refSize += encryption.KeyLength
-		chunkEncryption = newChunkEncryption(DefaultChunkSize, refSize)
+		chunkEncryption = newChunkEncryption(chunk.DefaultSize, refSize)
 	}
 
 	return &hasherStore{
@@ -190,9 +191,9 @@ func (h *hasherStore) decryptChunkData(chunkData ChunkData, encryptionKey encryp
 
 	// removing extra bytes which were just added for padding
 	length := ChunkData(decryptedSpan).Size()
-	for length > DefaultChunkSize {
-		length = length + (DefaultChunkSize - 1)
-		length = length / DefaultChunkSize
+	for length > chunk.DefaultSize {
+		length = length + (chunk.DefaultSize - 1)
+		length = length / chunk.DefaultSize
 		length *= h.refSize
 	}
 
diff --git a/swarm/storage/ldbstore_test.go b/swarm/storage/ldbstore_test.go
index baf9e8c14..5ee88baa5 100644
--- a/swarm/storage/ldbstore_test.go
+++ b/swarm/storage/ldbstore_test.go
@@ -27,6 +27,7 @@ import (
 	"time"
 
 	"github.com/ethereum/go-ethereum/common"
+	"github.com/ethereum/go-ethereum/swarm/chunk"
 	"github.com/ethereum/go-ethereum/swarm/log"
 	"github.com/ethereum/go-ethereum/swarm/storage/mock/mem"
 
@@ -184,7 +185,7 @@ func testIterator(t *testing.T, mock bool) {
 		t.Fatalf("init dbStore failed: %v", err)
 	}
 
-	chunks := GenerateRandomChunks(DefaultChunkSize, chunkcount)
+	chunks := GenerateRandomChunks(chunk.DefaultSize, chunkcount)
 
 	wg := &sync.WaitGroup{}
 	wg.Add(len(chunks))
@@ -294,7 +295,7 @@ func TestLDBStoreWithoutCollectGarbage(t *testing.T) {
 
 	chunks := []*Chunk{}
 	for i := 0; i < n; i++ {
-		c := GenerateRandomChunk(DefaultChunkSize)
+		c := GenerateRandomChunk(chunk.DefaultSize)
 		chunks = append(chunks, c)
 		log.Trace("generate random chunk", "idx", i, "chunk", c)
 	}
@@ -344,7 +345,7 @@ func TestLDBStoreCollectGarbage(t *testing.T) {
 
 	chunks := []*Chunk{}
 	for i := 0; i < n; i++ {
-		c := GenerateRandomChunk(DefaultChunkSize)
+		c := GenerateRandomChunk(chunk.DefaultSize)
 		chunks = append(chunks, c)
 		log.Trace("generate random chunk", "idx", i, "chunk", c)
 	}
@@ -398,7 +399,7 @@ func TestLDBStoreAddRemove(t *testing.T) {
 
 	chunks := []*Chunk{}
 	for i := 0; i < n; i++ {
-		c := GenerateRandomChunk(DefaultChunkSize)
+		c := GenerateRandomChunk(chunk.DefaultSize)
 		chunks = append(chunks, c)
 		log.Trace("generate random chunk", "idx", i, "chunk", c)
 	}
@@ -460,7 +461,7 @@ func TestLDBStoreRemoveThenCollectGarbage(t *testing.T) {
 
 	chunks := []*Chunk{}
 	for i := 0; i < capacity; i++ {
-		c := GenerateRandomChunk(DefaultChunkSize)
+		c := GenerateRandomChunk(chunk.DefaultSize)
 		chunks = append(chunks, c)
 		log.Trace("generate random chunk", "idx", i, "chunk", c)
 	}
diff --git a/swarm/storage/localstore.go b/swarm/storage/localstore.go
index 096d150ae..9e3474979 100644
--- a/swarm/storage/localstore.go
+++ b/swarm/storage/localstore.go
@@ -98,20 +98,16 @@ func NewTestLocalStoreForAddr(params *LocalStoreParams) (*LocalStore, error) {
 // After the LDBStore.Put, it is ensured that the MemStore
 // contains the chunk with the same data, but nil ReqC channel.
 func (ls *LocalStore) Put(ctx context.Context, chunk *Chunk) {
-	if l := len(chunk.SData); l < 9 {
-		log.Debug("incomplete chunk data", "addr", chunk.Addr, "length", l)
-		chunk.SetErrored(ErrChunkInvalid)
-		chunk.markAsStored()
-		return
-	}
 	valid := true
+	// ls.Validators contains a list of one validator per chunk type.
+	// if one validator succeeds, then the chunk is valid
 	for _, v := range ls.Validators {
 		if valid = v.Validate(chunk.Addr, chunk.SData); valid {
 			break
 		}
 	}
 	if !valid {
-		log.Trace("invalid content address", "addr", chunk.Addr)
+		log.Trace("invalid chunk", "addr", chunk.Addr, "len", len(chunk.SData))
 		chunk.SetErrored(ErrChunkInvalid)
 		chunk.markAsStored()
 		return
diff --git a/swarm/storage/localstore_test.go b/swarm/storage/localstore_test.go
index 2bb81efa3..ae62218fe 100644
--- a/swarm/storage/localstore_test.go
+++ b/swarm/storage/localstore_test.go
@@ -20,6 +20,8 @@ import (
 	"io/ioutil"
 	"os"
 	"testing"
+
+	"github.com/ethereum/go-ethereum/swarm/chunk"
 )
 
 var (
@@ -61,7 +63,7 @@ func TestValidator(t *testing.T) {
 	// add content address validator and check puts
 	// bad should fail, good should pass
 	store.Validators = append(store.Validators, NewContentAddressValidator(hashfunc))
-	chunks = GenerateRandomChunks(DefaultChunkSize, 2)
+	chunks = GenerateRandomChunks(chunk.DefaultSize, 2)
 	goodChunk = chunks[0]
 	badChunk = chunks[1]
 	copy(badChunk.SData, goodChunk.SData)
@@ -79,7 +81,7 @@ func TestValidator(t *testing.T) {
 	var negV boolTestValidator
 	store.Validators = append(store.Validators, negV)
 
-	chunks = GenerateRandomChunks(DefaultChunkSize, 2)
+	chunks = GenerateRandomChunks(chunk.DefaultSize, 2)
 	goodChunk = chunks[0]
 	badChunk = chunks[1]
 	copy(badChunk.SData, goodChunk.SData)
@@ -97,7 +99,7 @@ func TestValidator(t *testing.T) {
 	var posV boolTestValidator = true
 	store.Validators = append(store.Validators, posV)
 
-	chunks = GenerateRandomChunks(DefaultChunkSize, 2)
+	chunks = GenerateRandomChunks(chunk.DefaultSize, 2)
 	goodChunk = chunks[0]
 	badChunk = chunks[1]
 	copy(badChunk.SData, goodChunk.SData)
diff --git a/swarm/storage/mru/handler.go b/swarm/storage/mru/handler.go
index 32f43d502..57561fd14 100644
--- a/swarm/storage/mru/handler.go
+++ b/swarm/storage/mru/handler.go
@@ -21,17 +21,15 @@ package mru
 import (
 	"bytes"
 	"context"
-	"fmt"
 	"sync"
 	"time"
 	"unsafe"
 
+	"github.com/ethereum/go-ethereum/swarm/chunk"
 	"github.com/ethereum/go-ethereum/swarm/log"
 	"github.com/ethereum/go-ethereum/swarm/storage"
 )
 
-const chunkSize = 4096 // temporary until we implement FileStore in the resourcehandler
-
 type Handler struct {
 	chunkStore      *storage.NetStore
 	HashSize        int
@@ -66,8 +64,7 @@ func init() {
 }
 
 // NewHandler creates a new Mutable Resource API
-func NewHandler(params *HandlerParams) (*Handler, error) {
-
+func NewHandler(params *HandlerParams) *Handler {
 	rh := &Handler{
 		resources:       make(map[uint64]*resource),
 		storeTimeout:    defaultStoreTimeout,
@@ -82,7 +79,7 @@ func NewHandler(params *HandlerParams) (*Handler, error) {
 		hashPool.Put(hashfunc)
 	}
 
-	return rh, nil
+	return rh
 }
 
 // SetStore sets the store backend for the Mutable Resource API
@@ -94,9 +91,8 @@ func (h *Handler) SetStore(store *storage.NetStore) {
 // If it looks like a resource update, the chunk address is checked against the ownerAddr of the update's signature
 // It implements the storage.ChunkValidator interface
 func (h *Handler) Validate(chunkAddr storage.Address, data []byte) bool {
-
 	dataLength := len(data)
-	if dataLength < minimumChunkLength {
+	if dataLength < minimumChunkLength || dataLength > chunk.DefaultSize+8 {
 		return false
 	}
 
@@ -106,7 +102,7 @@ func (h *Handler) Validate(chunkAddr storage.Address, data []byte) bool {
 		rootAddr, _ := metadataHash(data)
 		valid := bytes.Equal(chunkAddr, rootAddr)
 		if !valid {
-			log.Debug(fmt.Sprintf("Invalid root metadata chunk with address: %s", chunkAddr.Hex()))
+			log.Debug("Invalid root metadata chunk with address", "addr", chunkAddr.Hex())
 		}
 		return valid
 	}
@@ -118,7 +114,7 @@ func (h *Handler) Validate(chunkAddr storage.Address, data []byte) bool {
 	// First, deserialize the chunk
 	var r SignedResourceUpdate
 	if err := r.fromChunk(chunkAddr, data); err != nil {
-		log.Debug("Invalid resource chunk with address %s: %s ", chunkAddr.Hex(), err.Error())
+		log.Debug("Invalid resource chunk", "addr", chunkAddr.Hex(), "err", err.Error())
 		return false
 	}
 
@@ -126,7 +122,7 @@ func (h *Handler) Validate(chunkAddr storage.Address, data []byte) bool {
 	// that was used to retrieve this chunk
 	// if this validation fails, someone forged a chunk.
 	if !bytes.Equal(chunkAddr, r.updateHeader.UpdateAddr()) {
-		log.Debug("period,version,rootAddr contained in update chunk do not match updateAddr %s", chunkAddr.Hex())
+		log.Debug("period,version,rootAddr contained in update chunk do not match updateAddr", "addr", chunkAddr.Hex())
 		return false
 	}
 
@@ -134,7 +130,7 @@ func (h *Handler) Validate(chunkAddr storage.Address, data []byte) bool {
 	// If it fails, it means either the signature is not valid, data is corrupted
 	// or someone is trying to update someone else's resource.
 	if err := r.Verify(); err != nil {
-		log.Debug("Invalid signature: %v", err)
+		log.Debug("Invalid signature", "err", err)
 		return false
 	}
 
@@ -172,11 +168,6 @@ func (h *Handler) GetVersion(rootAddr storage.Address) (uint32, error) {
 	return rsrc.version, nil
 }
 
-// \TODO should be hashsize * branches from the chosen chunker, implement with FileStore
-func (h *Handler) chunkSize() int64 {
-	return chunkSize
-}
-
 // New creates a new metadata chunk out of the request passed in.
 func (h *Handler) New(ctx context.Context, request *Request) error {
 
diff --git a/swarm/storage/mru/resource_test.go b/swarm/storage/mru/resource_test.go
index 95c9eccdf..76d7c58a1 100644
--- a/swarm/storage/mru/resource_test.go
+++ b/swarm/storage/mru/resource_test.go
@@ -30,6 +30,7 @@ import (
 	"github.com/ethereum/go-ethereum/contracts/ens"
 	"github.com/ethereum/go-ethereum/crypto"
 	"github.com/ethereum/go-ethereum/log"
+	"github.com/ethereum/go-ethereum/swarm/chunk"
 	"github.com/ethereum/go-ethereum/swarm/multihash"
 	"github.com/ethereum/go-ethereum/swarm/storage"
 )
@@ -776,14 +777,11 @@ func TestValidatorInStore(t *testing.T) {
 
 	// set up resource handler and add is as a validator to the localstore
 	rhParams := &HandlerParams{}
-	rh, err := NewHandler(rhParams)
-	if err != nil {
-		t.Fatal(err)
-	}
+	rh := NewHandler(rhParams)
 	store.Validators = append(store.Validators, rh)
 
 	// create content addressed chunks, one good, one faulty
-	chunks := storage.GenerateRandomChunks(storage.DefaultChunkSize, 2)
+	chunks := storage.GenerateRandomChunks(chunk.DefaultSize, 2)
 	goodChunk := chunks[0]
 	badChunk := chunks[1]
 	badChunk.SData = goodChunk.SData
diff --git a/swarm/storage/mru/testutil.go b/swarm/storage/mru/testutil.go
index 751f51af3..6efcba9ab 100644
--- a/swarm/storage/mru/testutil.go
+++ b/swarm/storage/mru/testutil.go
@@ -38,10 +38,7 @@ func (t *TestHandler) Close() {
 // NewTestHandler creates Handler object to be used for testing purposes.
 func NewTestHandler(datadir string, params *HandlerParams) (*TestHandler, error) {
 	path := filepath.Join(datadir, testDbDirName)
-	rh, err := NewHandler(params)
-	if err != nil {
-		return nil, fmt.Errorf("resource handler create fail: %v", err)
-	}
+	rh := NewHandler(params)
 	localstoreparams := storage.NewDefaultLocalStoreParams()
 	localstoreparams.Init(path)
 	localStore, err := storage.NewLocalStore(localstoreparams, nil)
diff --git a/swarm/storage/mru/update.go b/swarm/storage/mru/update.go
index 88c4ac4e5..d1bd37ddf 100644
--- a/swarm/storage/mru/update.go
+++ b/swarm/storage/mru/update.go
@@ -20,6 +20,7 @@ import (
 	"encoding/binary"
 	"errors"
 
+	"github.com/ethereum/go-ethereum/swarm/chunk"
 	"github.com/ethereum/go-ethereum/swarm/log"
 	"github.com/ethereum/go-ethereum/swarm/multihash"
 )
@@ -42,7 +43,7 @@ const chunkPrefixLength = 2 + 2
 //
 // Minimum size is Header + 1 (minimum data length, enforced)
 const minimumUpdateDataLength = updateHeaderLength + 1
-const maxUpdateDataLength = chunkSize - signatureLength - updateHeaderLength - chunkPrefixLength
+const maxUpdateDataLength = chunk.DefaultSize - signatureLength - updateHeaderLength - chunkPrefixLength
 
 // binaryPut serializes the resource update information into the given slice
 func (r *resourceUpdate) binaryPut(serializedData []byte) error {
diff --git a/swarm/storage/pyramid.go b/swarm/storage/pyramid.go
index 2923c81c5..36ff66d04 100644
--- a/swarm/storage/pyramid.go
+++ b/swarm/storage/pyramid.go
@@ -25,6 +25,7 @@ import (
 	"sync"
 	"time"
 
+	"github.com/ethereum/go-ethereum/swarm/chunk"
 	"github.com/ethereum/go-ethereum/swarm/log"
 )
 
@@ -101,11 +102,11 @@ func NewPyramidSplitterParams(addr Address, reader io.Reader, putter Putter, get
 	New chunks to store are store using the putter which the caller provides.
 */
 func PyramidSplit(ctx context.Context, reader io.Reader, putter Putter, getter Getter) (Address, func(context.Context) error, error) {
-	return NewPyramidSplitter(NewPyramidSplitterParams(nil, reader, putter, getter, DefaultChunkSize)).Split(ctx)
+	return NewPyramidSplitter(NewPyramidSplitterParams(nil, reader, putter, getter, chunk.DefaultSize)).Split(ctx)
 }
 
 func PyramidAppend(ctx context.Context, addr Address, reader io.Reader, putter Putter, getter Getter) (Address, func(context.Context) error, error) {
-	return NewPyramidSplitter(NewPyramidSplitterParams(addr, reader, putter, getter, DefaultChunkSize)).Append(ctx)
+	return NewPyramidSplitter(NewPyramidSplitterParams(addr, reader, putter, getter, chunk.DefaultSize)).Append(ctx)
 }
 
 // Entry to create a tree node
diff --git a/swarm/storage/types.go b/swarm/storage/types.go
index 3114ef576..53e3af485 100644
--- a/swarm/storage/types.go
+++ b/swarm/storage/types.go
@@ -30,6 +30,7 @@ import (
 	"github.com/ethereum/go-ethereum/common"
 	"github.com/ethereum/go-ethereum/crypto/sha3"
 	"github.com/ethereum/go-ethereum/swarm/bmt"
+	"github.com/ethereum/go-ethereum/swarm/chunk"
 )
 
 const MaxPO = 16
@@ -114,7 +115,9 @@ func MakeHashFunc(hash string) SwarmHasher {
 	case "BMT":
 		return func() SwarmHash {
 			hasher := sha3.NewKeccak256
-			pool := bmt.NewTreePool(hasher, bmt.SegmentCount, bmt.PoolSize)
+			hasherSize := hasher().Size()
+			segmentCount := chunk.DefaultSize / hasherSize
+			pool := bmt.NewTreePool(hasher, segmentCount, bmt.PoolSize)
 			return bmt.New(pool)
 		}
 	}
@@ -230,8 +233,8 @@ func GenerateRandomChunk(dataSize int64) *Chunk {
 func GenerateRandomChunks(dataSize int64, count int) (chunks []*Chunk) {
 	var i int
 	hasher := MakeHashFunc(DefaultHash)()
-	if dataSize > DefaultChunkSize {
-		dataSize = DefaultChunkSize
+	if dataSize > chunk.DefaultSize {
+		dataSize = chunk.DefaultSize
 	}
 
 	for i = 0; i < count; i++ {
@@ -345,6 +348,10 @@ func NewContentAddressValidator(hasher SwarmHasher) *ContentAddressValidator {
 
 // Validate that the given key is a valid content address for the given data
 func (v *ContentAddressValidator) Validate(addr Address, data []byte) bool {
+	if l := len(data); l < 9 || l > chunk.DefaultSize+8 {
+		return false
+	}
+
 	hasher := v.Hasher()
 	hasher.ResetWithLength(data[:8])
 	hasher.Write(data[8:])