Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Maybe the zkevm-node api of "TraceBlockByNumber" and "TraceBlockByHash" have a optimize for avoid elapsed time #3733

Open
lyh169 opened this issue Jul 18, 2024 · 0 comments

Comments

@lyh169
Copy link

lyh169 commented Jul 18, 2024

func (d *DebugEndpoints) TraceBlockByHash(hash types.ArgHash, cfg *traceConfig) (interface{}, types.Error) {
	return d.txMan.NewDbTxScope(d.state, func(ctx context.Context, dbTx pgx.Tx) (interface{}, types.Error) {
		block, err := d.state.GetL2BlockByHash(ctx, hash.Hash(), dbTx)
		if errors.Is(err, state.ErrNotFound) {
			return nil, types.NewRPCError(types.DefaultErrorCode, fmt.Sprintf("block %s not found", hash.Hash().String()))
		} else if err != nil {
			return RPCErrorResponse(types.DefaultErrorCode, "failed to get block by hash", err, true)
		}

		traces, rpcErr := d.buildTraceBlock(ctx, block.Transactions(), cfg, dbTx)
		if rpcErr != nil {
			return nil, rpcErr
		}

		return traces, nil
	})
}
func (d *DebugEndpoints) buildTraceBlock(ctx context.Context, txs []*ethTypes.Transaction, cfg *traceConfig, dbTx pgx.Tx) (interface{}, types.Error) {
	traces := []traceBlockTransactionResponse{}
	for _, tx := range txs {
		traceTransaction, err := d.buildTraceTransaction(ctx, tx.Hash(), cfg, dbTx)
		if err != nil {
			errMsg := fmt.Sprintf("failed to get trace for transaction %v: %v", tx.Hash().String(), err.Error())
			return RPCErrorResponse(types.DefaultErrorCode, errMsg, err, true)
		}
		traceBlockTransaction := traceBlockTransactionResponse{
			Result: traceTransaction,
		}
		traces = append(traces, traceBlockTransaction)
	}

	return traces, nil
}

The code of buildTraceBlock need build the trace transaction one by one; but the buildTraceTransaction need executor the tx from the first tx to the current index tx , so this do a lot of repetitive executor , like flower;

	for _, tx := range l2Block.Transactions() {
		checkReceipt, err := s.GetTransactionReceipt(ctx, tx.Hash(), dbTx)
		if err != nil {
			return nil, err
		}
		if checkReceipt.TransactionIndex < receipt.TransactionIndex {
			count++
		}
	}

	// since the executor only stores the state roots by block, we need to
	// execute all the txs in the block until the tx we want to trace
	var txsToEncode []types.Transaction
	var effectivePercentage []uint8
	for i := 0; i <= count; i++ {
		txsToEncode = append(txsToEncode, *l2Block.Transactions()[i])
		effectivePercentage = append(effectivePercentage, MaxEffectivePercentage)
		log.Debugf("trace will reprocess tx: %v", l2Block.Transactions()[i].Hash().String())
	}

But the go-ethereum traceBlock only need executor all tx of the block once;

func (api *API) traceBlock(ctx context.Context, block *types.Block, config *TraceConfig) ([]*txTraceResult, error) {
	if block.NumberU64() == 0 {
		return nil, errors.New("genesis is not traceable")
	}
	// Prepare base state
	parent, err := api.blockByNumberAndHash(ctx, rpc.BlockNumber(block.NumberU64()-1), block.ParentHash())
	if err != nil {
		return nil, err
	}
	reexec := defaultTraceReexec
	if config != nil && config.Reexec != nil {
		reexec = *config.Reexec
	}
	statedb, release, err := api.backend.StateAtBlock(ctx, parent, reexec, nil, true, false)
	if err != nil {
		return nil, err
	}
	defer release()
	// JS tracers have high overhead. In this case run a parallel
	// process that generates states in one thread and traces txes
	// in separate worker threads.
	if config != nil && config.Tracer != nil && *config.Tracer != "" {
		if isJS := DefaultDirectory.IsJS(*config.Tracer); isJS {
			return api.traceBlockParallel(ctx, block, statedb, config)
		}
	}
	// Native tracers have low overhead
	var (
		txs       = block.Transactions()
		blockHash = block.Hash()
		blockCtx  = core.NewEVMBlockContext(block.Header(), api.chainContext(ctx), nil)
		signer    = types.MakeSigner(api.backend.ChainConfig(), block.Number(), block.Time())
		results   = make([]*txTraceResult, len(txs))
	)
	if beaconRoot := block.BeaconRoot(); beaconRoot != nil {
		vmenv := vm.NewEVM(blockCtx, vm.TxContext{}, statedb, api.backend.ChainConfig(), vm.Config{})
		core.ProcessBeaconBlockRoot(*beaconRoot, vmenv, statedb)
	}
	for i, tx := range txs {
		// Generate the next state snapshot fast without tracing
		msg, _ := core.TransactionToMessage(tx, signer, block.BaseFee())
		txctx := &Context{
			BlockHash:   blockHash,
			BlockNumber: block.Number(),
			TxIndex:     i,
			TxHash:      tx.Hash(),
		}
		res, err := api.traceTx(ctx, tx, msg, txctx, blockCtx, statedb, config)
		if err != nil {
			return nil, err
		}
		results[i] = &txTraceResult{TxHash: tx.Hash(), Result: res}
	}
	return results, nil
}

Is zkevm-node have a optimizing strategy about this issue? Thanks

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant