1. 程式人生 > >以太坊原始碼分析(30)eth-bloombits和filter原始碼分析

以太坊原始碼分析(30)eth-bloombits和filter原始碼分析

## 以太坊的布隆過濾器
以太坊的區塊頭中包含了一個叫做logsBloom的區域。 這個區域儲存了當前區塊中所有的收據的日誌的布隆過濾器,一共是2048個bit。也就是256個位元組。
而我們的一個交易的收據包含了很多的日誌記錄。 每個日誌記錄包含了 合約的地址, 多個Topic。 而在我們的收據中也存在一個布隆過濾器,這個布隆過濾器記錄了所有的日誌記錄的資訊。如果我們看黃皮書裡面對日誌記錄的形式化定義。
O代表我們的日誌記錄,Oa代表logger的地址,Oto,Ot1代表日誌的Topics, Od代表時間。
Oa是20個位元組,Ot是32個位元組,Od是很多位元組
我們定義了一個布隆過濾器函式M,用來把一個日誌物件轉換成256位元組的hash
M3:2045是一個特別的函式,用來設定2048個bit位中的三位為1。
對於任意的輸入值,首先求他的KEC輸出, 然後通過取KEC輸出的 [0,1] [2,3],[4,5] 這幾位的值 對2048取模, 得到三個值, 這三個值就是輸出的2048中需要置位的下標。 也就是說對於任何一個輸入,如果它對應的三個下標的值不都為1,那麼它肯定不在這個區塊中。 當如如果對應的三位都為1,也不能說明一定在這個區塊中。 這就是布隆過濾器的特性。
收據中的布隆過濾器就是所有的日誌的布隆過濾器輸出的並集。
同時區塊頭中的logBloom,就是所有的收據的布隆過濾器的並集。
## ChainIndexer 和 BloomIndexer
最開始看到ChainIndexer,不是很明白是什麼功能。 其實從名字中可以看到,是Chain的索引。 在 eth中我們有看到BloomIndexer,這個就是布隆過濾器的索引。
在我們的協議中提供了查詢指定Log的功能。
使用者可以通過傳遞下面的引數來查詢指定的Log,開始的區塊號,結束的區塊號, 根據合約 Addresses指定的地址過濾,根據指定的Topics來過濾。
    // FilterCriteria represents a request to create a new filter.    type FilterCriteria struct {        FromBlock *big.Int        ToBlock *big.Int        Addresses []common.Address        Topics [][]common.Hash    }
如果開始和結束之間間隔很大,那麼如果直接依次檢索每個區塊頭的logBloom區域是比較低效的。 因為每個區塊頭都是分開儲存的, 可能需要非常多的磁碟隨機訪問。
所以以太坊協議在本地維護了一套索引,用來加速這個過程。
大致原理是。 每4096個區塊稱為一個Section,一個Section裡面的logBloom會儲存在一起。對於每個Section, 用一個二維資料,A[2048
][4096]來儲存。 第一維2048代表了bloom過濾器的長度2048個位元組。 第二維4096代表了一個Section裡面的所有區塊,每一個位置按照順序代表了其中的一個區塊。
- A[0][0]=blockchain[section*4096+0].logBloom[0],- A[0][1]=blockchain[section*4096+1].logBloom[0],- A[0][4096]=blockchain[section*4096+1].logBloom[0],- A[1][0]=blockchain[section*4096+0].logBloom[1],- A[1][1024]=blockchain[section*4096+1024].logBloom[1],-
A[2047][1]=blockchain[section*4096+1].logBloom[2047],
如果Section填充完畢,那麼會寫成2048個KV。 ![image](picture/bloom_6.png)

## bloombit.go 程式碼分析
這個程式碼相對不是很獨立,如果單獨看這個程式碼,有點摸不著頭腦的感覺, 因為它只是實現了一些介面,具體的處理邏輯並不在這裡,而是在core裡面。 不過這裡我先結合之前講到的資訊分析一下。 後續更詳細的邏輯在分析core的程式碼的時候再詳細分析。
服務執行緒startBloomHandlers,這個方法是為了響應具體的查詢請求, 給定指定的Section和bit來從levelDB裡面查詢然後返回出去。 單獨看這裡有點摸不著頭腦。 這個方法的呼叫比較複雜。 涉及到core裡面的很多邏輯。 這裡先不細說了。 直到有這個方法就行了。
    type Retrieval struct {        Bit uint           //Bit的取值 0-2047 代表了想要獲取哪一位的值        Sections []uint64       // 那些Section        Bitsets [][]byte       // 返回值 查詢出來的結果。    }    // startBloomHandlers starts a batch of goroutines to accept bloom bit database    // retrievals from possibly a range of filters and serving the data to satisfy.    func (eth *Ethereum) startBloomHandlers() {        for i := 0; i < bloomServiceThreads; i++ {            go func() {                for {                    select {                    case <-eth.shutdownChan:                        return                    case request := <-eth.bloomRequests: // request是一個通道                        task := <-request //從通道里面獲取一個task                        task.Bitsets = make([][]byte, len(task.Sections))                        for i, section := range task.Sections {                            head := core.GetCanonicalHash(eth.chainDb, (section+1)*params.BloomBitsBlocks-1)                            blob, err := bitutil.DecompressBytes(core.GetBloomBits(eth.chainDb, task.Bit, section, head), int(params.BloomBitsBlocks)/8)                            if err != nil {                                panic(err)                            }                            task.Bitsets[i] = blob                        }                        request <- task //通過request通道返回結果                    }                }            }()        }    }

### 資料結構BloomIndexer物件主要使用者構建索引的過程,是core.ChainIndexer的一個介面實現,所以只實現了一些必須的介面。對於建立索引的邏輯還在core.ChainIndexer裡面。

    // BloomIndexer implements a core.ChainIndexer, building up a rotated bloom bits index    // for the Ethereum header bloom filters, permitting blazing fast filtering.    type BloomIndexer struct {        size uint64 // section size to generate bloombits for        db ethdb.Database // database instance to write index data and metadata into        gen *bloombits.Generator // generator to rotate the bloom bits crating the bloom index        section uint64 // Section is the section number being processed currently 當前的section        head common.Hash // Head is the hash of the last header processed    }
    // NewBloomIndexer returns a chain indexer that generates bloom bits data for the    // canonical chain for fast logs filtering.    func NewBloomIndexer(db ethdb.Database, size uint64) *core.ChainIndexer {        backend := &BloomIndexer{            db: db,            size: size,        }        table := ethdb.NewTable(db, string(core.BloomBitsIndexPrefix))        return core.NewChainIndexer(db, table, backend, size, bloomConfirms, bloomThrottling, "bloombits")    }
Reset實現了ChainIndexerBackend的方法,啟動一個新的section
    // Reset implements core.ChainIndexerBackend, starting a new bloombits index    // section.    func (b *BloomIndexer) Reset(section uint64) {        gen, err := bloombits.NewGenerator(uint(b.size))        if err != nil {            panic(err)        }        b.gen, b.section, b.head = gen, section, common.Hash{}    }
Process實現了ChainIndexerBackend, 增加一個新的區塊頭到index    // Process implements core.ChainIndexerBackend, adding a new header's bloom into    // the index.    func (b *BloomIndexer) Process(header *types.Header) {        b.gen.AddBloom(uint(header.Number.Uint64()-b.section*b.size), header.Bloom)        b.head = header.Hash()    }
Commit方法實現了ChainIndexerBackend,持久化並寫入資料庫。    // Commit implements core.ChainIndexerBackend, finalizing the bloom section and    // writing it out into the database.    func (b *BloomIndexer) Commit() error {        batch := b.db.NewBatch()        for i := 0; i < types.BloomBitLength; i++ {            bits, err := b.gen.Bitset(uint(i))            if err != nil {                return err            }            core.WriteBloomBits(batch, uint(i), b.section, b.head, bitutil.CompressBytes(bits))        }        return batch.Write()    }
## filter/api.go 原始碼分析
eth/filter 包 包含了給使用者提供過濾的功能,使用者可以通過呼叫對交易或者區塊進行過濾,然後持續的獲取結果,如果5分鐘沒有操作,這個過濾器會被刪除。

過濾器的結構。
    var (        deadline = 5 * time.Minute // consider a filter inactive if it has not been polled for within deadline    )    // filter is a helper struct that holds meta information over the filter type    // and associated subscription in the event system.    type filter struct {        typ Type           // 過濾器的型別, 過濾什麼型別的資料        deadline *time.Timer // filter is inactiv when deadline triggers 當計時器響起的時候,會觸發定時器。        hashes []common.Hash //過濾出來的hash結果        crit FilterCriteria //過濾條件        logs []*types.Log //過濾出來的Log資訊        s *Subscription // associated subscription in event system 事件系統中的訂閱器。    }
構造方法
    // PublicFilterAPI offers support to create and manage filters. This will allow external clients to retrieve various    // information related to the Ethereum protocol such als blocks, transactions and logs.    // PublicFilterAPI用來建立和管理過濾器。 允許外部的客戶端獲取以太坊協議的一些資訊,比如區塊資訊,交易資訊和日誌資訊。    type PublicFilterAPI struct {        backend Backend        mux *event.TypeMux        quit chan struct{}        chainDb ethdb.Database        events *EventSystem        filtersMu sync.Mutex        filters map[rpc.ID]*filter    }    // NewPublicFilterAPI returns a new PublicFilterAPI instance.    func NewPublicFilterAPI(backend Backend, lightMode bool) *PublicFilterAPI {        api := &PublicFilterAPI{            backend: backend,            mux: backend.EventMux(),            chainDb: backend.ChainDb(),            events: NewEventSystem(backend.EventMux(), backend, lightMode),            filters: make(map[rpc.ID]*filter),        }        go api.timeoutLoop()        return api    }
### 超時檢查    // timeoutLoop runs every 5 minutes and deletes filters that have not been recently used.    // Tt is started when the api is created.    // 每隔5分鐘檢查一下。 如果過期的過濾器,刪除。    func (api *PublicFilterAPI) timeoutLoop() {        ticker := time.NewTicker(5 * time.Minute)        for {            <-ticker.C            api.filtersMu.Lock()            for id, f := range api.filters {                select {                case <-f.deadline.C:                    f.s.Unsubscribe()                    delete(api.filters, id)                default:                    continue                }            }            api.filtersMu.Unlock()        }    }


NewPendingTransactionFilter,用來建立一個PendingTransactionFilter。 這種方式是用來給那種無法建立長連線的通道使用的(比如HTTP), 如果對於可以建立長連結的通道(比如WebSocket)可以使用rpc提供的傳送訂閱模式來處理,就不用持續的輪詢了    // NewPendingTransactionFilter creates a filter that fetches pending transaction hashes    // as transactions enter the pending state.    //    // It is part of the filter package because this filter can be used throug the    // `eth_getFilterChanges` polling method that is also used for log filters.    //    // https://github.com/ethereum/wiki/wiki/JSON-RPC#eth_newpendingtransactionfilter    func (api *PublicFilterAPI) NewPendingTransactionFilter() rpc.ID {        var (            pendingTxs = make(chan common.Hash)            // 在事件系統訂閱這種訊息            pendingTxSub = api.events.SubscribePendingTxEvents(pendingTxs)        )        api.filtersMu.Lock()        api.filters[pendingTxSub.ID] = &filter{typ: PendingTransactionsSubscription, deadline: time.NewTimer(deadline), hashes: make([]common.Hash, 0), s: pendingTxSub}        api.filtersMu.Unlock()        go func() {            for {                select {                case ph := <-pendingTxs: // 接收到pendingTxs,儲存在過濾器的hashes容器裡面。                    api.filtersMu.Lock()                    if f, found := api.filters[pendingTxSub.ID]; found {                        f.hashes = append(f.hashes, ph)                    }                    api.filtersMu.Unlock()                case <-pendingTxSub.Err():                    api.filtersMu.Lock()                    delete(api.filters, pendingTxSub.ID)                    api.filtersMu.Unlock()                    return                }            }        }()        return pendingTxSub.ID    }
輪詢: GetFilterChanges    // GetFilterChanges returns the logs for the filter with the given id since    // last time it was called. This can be used for polling.    // GetFilterChanges 用來返回從上次呼叫到現在的所有的指定id的所有過濾資訊。這個可以用來輪詢。    // For pending transaction and block filters the result is []common.Hash.    // (pending)Log filters return []Log.    // 對於pending transaction和block的過濾器,返回結果型別是[]common.Hash. 對於pending Log 過濾器,返回的是 []Log    // https://github.com/ethereum/wiki/wiki/JSON-RPC#eth_getfilterchanges    func (api *PublicFilterAPI) GetFilterChanges(id rpc.ID) (interface{}, error) {        api.filtersMu.Lock()        defer api.filtersMu.Unlock()        if f, found := api.filters[id]; found {            if !f.deadline.Stop() { // 如果定時器已經觸發,但是filter還沒有移除,那麼我們先接收定時器的值,然後重置定時器                // timer expired but filter is not yet removed in timeout loop                // receive timer value and reset timer                <-f.deadline.C            }            f.deadline.Reset(deadline)            switch f.typ {            case PendingTransactionsSubscription, BlocksSubscription:                hashes := f.hashes                f.hashes = nil                return returnHashes(hashes), nil            case LogsSubscription:                logs := f.logs                f.logs = nil                return returnLogs(logs), nil            }        }        return []interface{}{}, fmt.Errorf("filter not found")    }


對於可以建立長連線的通道,可以直接使用rpc的傳送訂閱模式, 這樣客戶端就可以直接接收到過濾資訊,不用呼叫輪詢的方式了。 可以看到這種模式下面並沒有新增到filters這個容器,也沒有超時管理了。也就是說支援兩種模式。
    // NewPendingTransactions creates a subscription that is triggered each time a transaction    // enters the transaction pool and was signed from one of the transactions this nodes manages.    func (api *PublicFilterAPI) NewPendingTransactions(ctx context.Context) (*rpc.Subscription, error) {        notifier, supported := rpc.NotifierFromContext(ctx)        if !supported {            return &rpc.Subscription{}, rpc.ErrNotificationsUnsupported        }        rpcSub := notifier.CreateSubscription()        go func() {            txHashes := make(chan common.Hash)            pendingTxSub := api.events.SubscribePendingTxEvents(txHashes)            for {                select {                case h := <-txHashes:                    notifier.Notify(rpcSub.ID, h)                case <-rpcSub.Err():                    pendingTxSub.Unsubscribe()                    return                case <-notifier.Closed():                    pendingTxSub.Unsubscribe()                    return                }            }        }()        return rpcSub, nil    }

日誌過濾功能,根據FilterCriteria指定的引數,來對日誌進行過濾,開始區塊,結束區塊,地址和Topics,這裡面引入了一個新的物件filter    // FilterCriteria represents a request to create a new filter.    type FilterCriteria struct {        FromBlock *big.Int        ToBlock *big.Int        Addresses []common.Address        Topics [][]common.Hash    }    // GetLogs returns logs matching the given argument that are stored within the state.    //    // https://github.com/ethereum/wiki/wiki/JSON-RPC#eth_getlogs    func (api *PublicFilterAPI) GetLogs(ctx context.Context, crit FilterCriteria) ([]*types.Log, error) {        // Convert the RPC block numbers into internal representations        if crit.FromBlock == nil {            crit.FromBlock = big.NewInt(rpc.LatestBlockNumber.Int64())        }        if crit.ToBlock == nil {            crit.ToBlock = big.NewInt(rpc.LatestBlockNumber.Int64())        }        // Create and run the filter to get all the logs        // 建立了一個Filter物件 然後呼叫filter.Logs        filter := New(api.backend, crit.FromBlock.Int64(), crit.ToBlock.Int64(), crit.Addresses, crit.Topics)        logs, err := filter.Logs(ctx)        if err != nil {            return nil, err        }        return returnLogs(logs), err    }

## filter.gofiter.go裡面定義了一個Filter物件。這個物件主要用來根據 區塊的BloomIndexer和布隆過濾器等來執行日誌的過濾功能。
### 資料結構    // 後端, 這個後端其實是在core裡面實現的。 布隆過濾器的主要演算法在core裡面實現了。    type Backend interface {        ChainDb() ethdb.Database        EventMux() *event.TypeMux        HeaderByNumber(ctx context.Context, blockNr rpc.BlockNumber) (*types.Header, error)        GetReceipts(ctx context.Context, blockHash common.Hash) (types.Receipts, error)        SubscribeTxPreEvent(chan<- core.TxPreEvent) event.Subscription        SubscribeChainEvent(ch chan<- core.ChainEvent) event.Subscription        SubscribeRemovedLogsEvent(ch chan<- core.RemovedLogsEvent) event.Subscription        SubscribeLogsEvent(ch chan<- []*types.Log) event.Subscription        BloomStatus() (uint64, uint64)        ServiceFilter(ctx context.Context, session *bloombits.MatcherSession)    }    // Filter can be used to retrieve and filter logs.    type Filter struct {        backend Backend             // 後端        db ethdb.Database   // 資料庫        begin, end int64            // 開始結束區塊        addresses []common.Address // 篩選地址        topics [][]common.Hash  // 篩選主題        matcher *bloombits.Matcher  // 布隆過濾器的匹配器    }
建構函式把address和topic都加入到filters容器。然後構建了一個bloombits.NewMatcher(size, filters)。這個函式在core裡面實現, 暫時不會講解。
    // New creates a new filter which uses a bloom filter on blocks to figure out whether    // a particular block is interesting or not.    func New(backend Backend, begin, end int64, addresses []common.Address, topics [][]common.Hash) *Filter {        // Flatten the address and topic filter clauses into a single bloombits filter        // system. Since the bloombits are not positional, nil topics are permitted,        // which get flattened into a nil byte slice.        var filters [][][]byte        if len(addresses) > 0 {            filter := make([][]byte, len(addresses))            for i, address := range addresses {                filter[i] = address.Bytes()            }            filters = append(filters, filter)        }        for _, topicList := range topics {            filter := make([][]byte, len(topicList))            for i, topic := range topicList {                filter[i] = topic.Bytes()            }            filters = append(filters, filter)        }        // Assemble and return the filter        size, _ := backend.BloomStatus()        return &Filter{            backend: backend,            begin: begin,            end: end,            addresses: addresses,            topics: topics,            db: backend.ChainDb(),            matcher: bloombits.NewMatcher(size, filters),        }    }

Logs 執行過濾
    // Logs searches the blockchain for matching log entries, returning all from the    // first block that contains matches, updating the start of the filter accordingly.    func (f *Filter) Logs(ctx context.Context) ([]*types.Log, error) {        // Figure out the limits of the filter range        header, _ := f.backend.HeaderByNumber(ctx, rpc.LatestBlockNumber)        if header == nil {            return nil, nil        }        head := header.Number.Uint64()        if f.begin == -1 {            f.begin = int64(head)        }        end := uint64(f.end)        if f.end == -1 {            end = head        }        // Gather all indexed logs, and finish with non indexed ones        var (            logs []*types.Log            err error        )        size, sections := f.backend.BloomStatus()        // indexed 是指建立了索引的區塊的最大值。 如果過濾的範圍落在了建立了索引的部分。        // 那麼執行索引搜尋。        if indexed := sections * size; indexed > uint64(f.begin) {            if indexed > end {                logs, err = f.indexedLogs(ctx, end)            } else {                logs, err = f.indexedLogs(ctx, indexed-1)            }            if err != nil {                return logs, err            }        }        // 對於剩下的部分執行非索引的搜尋。        rest, err := f.unindexedLogs(ctx, end)        logs = append(logs, rest...)        return logs, err    }

索引搜尋
    // indexedLogs returns the logs matching the filter criteria based on the bloom    // bits indexed available locally or via the network.    func (f *Filter) indexedLogs(ctx context.Context, end uint64) ([]*types.Log, error) {        // Create a matcher session and request servicing from the backend        matches := make(chan uint64, 64)        // 啟動matcher        session, err := f.matcher.Start(uint64(f.begin), end, matches)        if err != nil {            return nil, err        }        defer session.Close(time.Second)        // 進行過濾服務。 這些都在core裡面。後續分析core的程式碼會進行分析。        f.backend.ServiceFilter(ctx, session)        // Iterate over the matches until exhausted or context closed        var logs []*types.Log        for {            select {            case number, ok := <-matches:                // Abort if all matches have been fulfilled                if !ok { // 沒有接收到值並且channel已經被關閉                    f.begin = int64(end) + 1 //更新begin。以便於下面的非索引搜尋                    return logs, nil                }                // Retrieve the suggested block and pull any truly matching logs                header, err := f.backend.HeaderByNumber(ctx, rpc.BlockNumber(number))                if header == nil || err != nil {                    return logs, err                }                found, err := f.checkMatches(ctx, header) //查詢匹配的值                if err != nil {                    return logs, err                }                logs = append(logs, found...)            case <-ctx.Done():                return logs, ctx.Err()            }        }    }
checkMatches,拿到所有的收據,並從收據中拿到所有的日誌。 執行filterLogs方法。    // checkMatches checks if the receipts belonging to the given header contain any log events that    // match the filter criteria. This function is called when the bloom filter signals a potential match.    func (f *Filter) checkMatches(ctx context.Context, header *types.Header) (logs []*types.Log, err error) {        // Get the logs of the block        receipts, err := f.backend.GetReceipts(ctx, header.Hash())        if err != nil {            return nil, err        }        var unfiltered []*types.Log        for _, receipt := range receipts {            unfiltered = append(unfiltered, ([]*types.Log)(receipt.Logs)...)        }        logs = filterLogs(unfiltered, nil, nil, f.addresses, f.topics)        if len(logs) > 0 {            return logs, nil        }        return nil, nil    }
filterLogs,這個方法從給定的logs裡面找到能夠匹配上的。並返回。
    // filterLogs creates a slice of logs matching the given criteria.    func filterLogs(logs []*types.Log, fromBlock, toBlock *big.Int, addresses []common.Address, topics [][]common.Hash) []*types.Log {        var ret []*types.Log    Logs:        for _, log := range logs {            if fromBlock != nil && fromBlock.Int64() >= 0 && fromBlock.Uint64() > log.BlockNumber {                continue            }            if toBlock != nil && toBlock.Int64() >= 0 && toBlock.Uint64() < log.BlockNumber {                continue            }            if len(addresses) > 0 && !includes(addresses, log.Address) {                continue            }            // If the to filtered topics is greater than the amount of topics in logs, skip.            if len(topics) > len(log.Topics) {                continue Logs            }            for i, topics := range topics {                match := len(topics) == 0 // empty rule set == wildcard                for _, topic := range topics {                    if log.Topics[i] == topic {                        match = true                        break                    }                }                if !match {                    continue Logs                }            }            ret = append(ret, log)        }        return ret    }
unindexedLogs,非索引查詢,迴圈遍歷所有的區塊。 首先用區塊裡面的header.Bloom來看是否有可能存在,如果有可能存在, 再使用checkMatches來檢索所有的匹配。    // indexedLogs returns the logs matching the filter criteria based on raw block    // iteration and bloom matching.    func (f *Filter) unindexedLogs(ctx context.Context, end uint64) ([]*types.Log, error) {        var logs []*types.Log        for ; f.begin <= int64(end); f.begin++ {            header, err := f.backend.HeaderByNumber(ctx, rpc.BlockNumber(f.begin))            if header == nil || err != nil {                return logs, err            }            if bloomFilter(header.Bloom, f.addresses, f.topics) {                found, err := f.checkMatches(ctx, header)                if err != nil {                    return logs, err                }                logs = append(logs, found...)            }        }        return logs, nil    }
## 總結filter原始碼包主要實現了兩個功能, - 提供了 釋出訂閱模式的filter RPC。用來給rpc客戶端提供實時的交易,區塊,日誌等的過濾- 提供了 基於bloomIndexer的日誌過濾模式,這種模式下,可以快速的對大量區塊執行布隆過濾操作。 還提供了歷史的日誌的過濾操作。




網址:http://www.qukuailianxueyuan.io/



欲領取造幣技術與全套虛擬機器資料

區塊鏈技術交流QQ群:756146052  備註:CSDN

尹成學院微信:備註:CSDN