lbcd/mempool/mempool.go

1554 lines
53 KiB
Go
Raw Normal View History

2016-02-12 19:44:56 +01:00
// Copyright (c) 2013-2016 The btcsuite developers
// Use of this source code is governed by an ISC
// license that can be found in the LICENSE file.
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
package mempool
import (
"container/list"
"fmt"
2014-07-02 15:50:08 +02:00
"math"
"sync"
2016-02-12 19:44:56 +01:00
"sync/atomic"
2014-07-02 15:50:08 +02:00
"time"
"github.com/btcsuite/btcd/blockchain"
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
"github.com/btcsuite/btcd/blockchain/indexers"
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
"github.com/btcsuite/btcd/btcjson"
"github.com/btcsuite/btcd/chaincfg"
"github.com/btcsuite/btcd/chaincfg/chainhash"
"github.com/btcsuite/btcd/mining"
"github.com/btcsuite/btcd/txscript"
"github.com/btcsuite/btcd/wire"
"github.com/btcsuite/btcutil"
)
const (
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
// DefaultBlockPrioritySize is the default size in bytes for high-
// priority / low-fee transactions. It is used to help determine which
// are allowed into the mempool and consequently affects their relay and
// inclusion when generating block templates.
DefaultBlockPrioritySize = 50000
// orphanTTL is the maximum amount of time an orphan is allowed to
// stay in the orphan pool before it expires and is evicted during the
// next scan.
orphanTTL = time.Minute * 15
// orphanExpireScanInterval is the minimum amount of time in between
// scans of the orphan pool to evict expired transactions.
orphanExpireScanInterval = time.Minute * 5
// MaxRBFSequence is the maximum sequence number an input can use to
// signal that the transaction spending it can be replaced using the
// Replace-By-Fee (RBF) policy.
MaxRBFSequence = 0xfffffffd
// MaxReplacementEvictions is the maximum number of transactions that
// can be evicted from the mempool when accepting a transaction
// replacement.
MaxReplacementEvictions = 100
)
// Tag represents an identifier to use for tagging orphan transactions. The
// caller may choose any scheme it desires, however it is common to use peer IDs
// so that orphans can be identified by which peer first relayed them.
type Tag uint64
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
// Config is a descriptor containing the memory pool configuration.
type Config struct {
// Policy defines the various mempool configuration options related
// to policy.
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
Policy Policy
// ChainParams identifies which chain parameters the txpool is
// associated with.
ChainParams *chaincfg.Params
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// FetchUtxoView defines the function to use to fetch unspent
// transaction output information.
FetchUtxoView func(*btcutil.Tx) (*blockchain.UtxoViewpoint, error)
// BestHeight defines the function to use to access the block height of
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// the current best chain.
BestHeight func() int32
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// MedianTimePast defines the function to use in order to access the
// median time past calculated from the point-of-view of the current
// chain tip within the best chain.
MedianTimePast func() time.Time
// CalcSequenceLock defines the function to use in order to generate
// the current sequence lock for the given transaction using the passed
// utxo view.
CalcSequenceLock func(*btcutil.Tx, *blockchain.UtxoViewpoint) (*blockchain.SequenceLock, error)
// IsDeploymentActive returns true if the target deploymentID is
// active, and false otherwise. The mempool uses this function to gauge
// if transactions using new to be soft-forked rules should be allowed
// into the mempool or not.
IsDeploymentActive func(deploymentID uint32) (bool, error)
// SigCache defines a signature cache to use.
SigCache *txscript.SigCache
// HashCache defines the transaction hash mid-state cache to use.
HashCache *txscript.HashCache
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
// AddrIndex defines the optional address index instance to use for
// indexing the unconfirmed transactions in the memory pool.
// This can be nil if the address index is not enabled.
AddrIndex *indexers.AddrIndex
// FeeEstimatator provides a feeEstimator. If it is not nil, the mempool
// records all new transactions it observes into the feeEstimator.
FeeEstimator *FeeEstimator
}
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
// Policy houses the policy (configuration parameters) which is used to
// control the mempool.
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
type Policy struct {
// MaxTxVersion is the transaction version that the mempool should
// accept. All transactions above this version are rejected as
// non-standard.
MaxTxVersion int32
// DisableRelayPriority defines whether to relay free or low-fee
// transactions that do not have enough priority to be relayed.
DisableRelayPriority bool
// AcceptNonStd defines whether to accept non-standard transactions. If
// true, non-standard transactions will be accepted into the mempool.
// Otherwise, all non-standard transactions will be rejected.
AcceptNonStd bool
// FreeTxRelayLimit defines the given amount in thousands of bytes
// per minute that transactions with no fee are rate limited to.
FreeTxRelayLimit float64
// MaxOrphanTxs is the maximum number of orphan transactions
// that can be queued.
MaxOrphanTxs int
// MaxOrphanTxSize is the maximum size allowed for orphan transactions.
// This helps prevent memory exhaustion attacks from sending a lot of
// of big orphans.
MaxOrphanTxSize int
// MaxSigOpCostPerTx is the cumulative maximum cost of all the signature
// operations in a single transaction we will relay or mine. It is a
// fraction of the max signature operations for a block.
MaxSigOpCostPerTx int
// MinRelayTxFee defines the minimum transaction fee in BTC/kB to be
// considered a non-zero fee.
MinRelayTxFee btcutil.Amount
// RejectReplacement, if true, rejects accepting replacement
// transactions using the Replace-By-Fee (RBF) signaling policy into
// the mempool.
RejectReplacement bool
}
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
// TxDesc is a descriptor containing a transaction in the mempool along with
// additional metadata.
type TxDesc struct {
mining.TxDesc
// StartingPriority is the priority of the transaction when it was added
// to the pool.
StartingPriority float64
}
// orphanTx is normal transaction that references an ancestor transaction
// that is not yet available. It also contains additional information related
// to it such as an expiration time to help prevent caching the orphan forever.
type orphanTx struct {
tx *btcutil.Tx
tag Tag
expiration time.Time
}
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
// TxPool is used as a source of transactions that need to be mined into blocks
// and relayed to other peers. It is safe for concurrent access from multiple
// peers.
type TxPool struct {
2016-02-12 19:44:56 +01:00
// The following variables must only be used atomically.
lastUpdated int64 // last time pool was updated
mtx sync.RWMutex
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
cfg Config
pool map[chainhash.Hash]*TxDesc
orphans map[chainhash.Hash]*orphanTx
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
orphansByPrev map[wire.OutPoint]map[chainhash.Hash]*btcutil.Tx
outpoints map[wire.OutPoint]*btcutil.Tx
2016-02-12 19:44:56 +01:00
pennyTotal float64 // exponentially decaying total for penny spends.
lastPennyUnix int64 // unix time of last ``penny spend''
// nextExpireScan is the time after which the orphan pool will be
// scanned in order to evict orphans. This is NOT a hard deadline as
// the scan will only run when an orphan is added to the pool as opposed
// to on an unconditional timer.
nextExpireScan time.Time
}
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
// Ensure the TxPool type implements the mining.TxSource interface.
var _ mining.TxSource = (*TxPool)(nil)
// removeOrphan is the internal function which implements the public
// RemoveOrphan. See the comment for RemoveOrphan for more details.
//
// This function MUST be called with the mempool lock held (for writes).
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
func (mp *TxPool) removeOrphan(tx *btcutil.Tx, removeRedeemers bool) {
// Nothing to do if passed tx is not an orphan.
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
txHash := tx.Hash()
otx, exists := mp.orphans[*txHash]
if !exists {
return
}
// Remove the reference from the previous orphan index.
for _, txIn := range otx.tx.MsgTx().TxIn {
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
orphans, exists := mp.orphansByPrev[txIn.PreviousOutPoint]
if exists {
delete(orphans, *txHash)
// Remove the map entry altogether if there are no
// longer any orphans which depend on it.
if len(orphans) == 0 {
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
delete(mp.orphansByPrev, txIn.PreviousOutPoint)
}
}
}
// Remove any orphans that redeem outputs from this one if requested.
if removeRedeemers {
prevOut := wire.OutPoint{Hash: *txHash}
for txOutIdx := range tx.MsgTx().TxOut {
prevOut.Index = uint32(txOutIdx)
for _, orphan := range mp.orphansByPrev[prevOut] {
mp.removeOrphan(orphan, true)
}
}
}
// Remove the transaction from the orphan pool.
delete(mp.orphans, *txHash)
}
// RemoveOrphan removes the passed orphan transaction from the orphan pool and
// previous orphan index.
//
// This function is safe for concurrent access.
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
func (mp *TxPool) RemoveOrphan(tx *btcutil.Tx) {
mp.mtx.Lock()
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
mp.removeOrphan(tx, false)
mp.mtx.Unlock()
}
// RemoveOrphansByTag removes all orphan transactions tagged with the provided
// identifier.
//
// This function is safe for concurrent access.
func (mp *TxPool) RemoveOrphansByTag(tag Tag) uint64 {
var numEvicted uint64
mp.mtx.Lock()
for _, otx := range mp.orphans {
if otx.tag == tag {
mp.removeOrphan(otx.tx, true)
numEvicted++
}
}
mp.mtx.Unlock()
return numEvicted
}
// limitNumOrphans limits the number of orphan transactions by evicting a random
// orphan if adding a new one would cause it to overflow the max allowed.
//
// This function MUST be called with the mempool lock held (for writes).
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
func (mp *TxPool) limitNumOrphans() error {
// Scan through the orphan pool and remove any expired orphans when it's
// time. This is done for efficiency so the scan only happens
// periodically instead of on every orphan added to the pool.
if now := time.Now(); now.After(mp.nextExpireScan) {
origNumOrphans := len(mp.orphans)
for _, otx := range mp.orphans {
if now.After(otx.expiration) {
// Remove redeemers too because the missing
// parents are very unlikely to ever materialize
// since the orphan has already been around more
// than long enough for them to be delivered.
mp.removeOrphan(otx.tx, true)
}
}
// Set next expiration scan to occur after the scan interval.
mp.nextExpireScan = now.Add(orphanExpireScanInterval)
numOrphans := len(mp.orphans)
if numExpired := origNumOrphans - numOrphans; numExpired > 0 {
log.Debugf("Expired %d %s (remaining: %d)", numExpired,
pickNoun(numExpired, "orphan", "orphans"),
numOrphans)
}
}
// Nothing to do if adding another orphan will not cause the pool to
// exceed the limit.
mempool: Optimize orphan map limiting. This optimizes the way in which the mempool oprhan map is limited in the same way the server block manager maps were previously optimized. Previously the code would read a cryptographically random value large enough to construct a hash, find the first entry larger than that value, and evict it. That approach is quite inefficient and could easily become a bottleneck when processing transactions due to the need to read from a source such as /dev/urandom and all of the subsequent hash comparisons. Luckily, strong cryptographic randomness is not needed here. The primary intent of limiting the maps is to control memory usage with a secondary concern of making it difficult for adversaries to force eviction of specific entries. Consequently, this changes the code to make use of the pseudorandom iteration order of Go's maps along with the preimage resistance of the hashing function to provide the desired functionality. It has previously been discussed that the specific pseudorandom iteration order is not guaranteed by the Go spec even though in practice that is how it is implemented. This is not a concern however because even if the specific compiler doesn't implement that, the preimage resistance of the hashing function alone is enough. The following is a before and after comparison of the function for both speed and memory allocations: benchmark old ns/op new ns/op delta ---------------------------------------------------------------- BenchmarkLimitNumOrphans 3727 243 -93.48% benchmark old allocs new allocs delta ----------------------------------------------------------------- BenchmarkLimitNumOrphans 4 0 -100.00%
2016-10-24 04:14:09 +02:00
if len(mp.orphans)+1 <= mp.cfg.Policy.MaxOrphanTxs {
return nil
}
mempool: Optimize orphan map limiting. This optimizes the way in which the mempool oprhan map is limited in the same way the server block manager maps were previously optimized. Previously the code would read a cryptographically random value large enough to construct a hash, find the first entry larger than that value, and evict it. That approach is quite inefficient and could easily become a bottleneck when processing transactions due to the need to read from a source such as /dev/urandom and all of the subsequent hash comparisons. Luckily, strong cryptographic randomness is not needed here. The primary intent of limiting the maps is to control memory usage with a secondary concern of making it difficult for adversaries to force eviction of specific entries. Consequently, this changes the code to make use of the pseudorandom iteration order of Go's maps along with the preimage resistance of the hashing function to provide the desired functionality. It has previously been discussed that the specific pseudorandom iteration order is not guaranteed by the Go spec even though in practice that is how it is implemented. This is not a concern however because even if the specific compiler doesn't implement that, the preimage resistance of the hashing function alone is enough. The following is a before and after comparison of the function for both speed and memory allocations: benchmark old ns/op new ns/op delta ---------------------------------------------------------------- BenchmarkLimitNumOrphans 3727 243 -93.48% benchmark old allocs new allocs delta ----------------------------------------------------------------- BenchmarkLimitNumOrphans 4 0 -100.00%
2016-10-24 04:14:09 +02:00
// Remove a random entry from the map. For most compilers, Go's
// range statement iterates starting at a random item although
// that is not 100% guaranteed by the spec. The iteration order
// is not important here because an adversary would have to be
// able to pull off preimage attacks on the hashing function in
// order to target eviction of specific entries anyways.
for _, otx := range mp.orphans {
// Don't remove redeemers in the case of a random eviction since
// it is quite possible it might be needed again shortly.
mp.removeOrphan(otx.tx, false)
mempool: Optimize orphan map limiting. This optimizes the way in which the mempool oprhan map is limited in the same way the server block manager maps were previously optimized. Previously the code would read a cryptographically random value large enough to construct a hash, find the first entry larger than that value, and evict it. That approach is quite inefficient and could easily become a bottleneck when processing transactions due to the need to read from a source such as /dev/urandom and all of the subsequent hash comparisons. Luckily, strong cryptographic randomness is not needed here. The primary intent of limiting the maps is to control memory usage with a secondary concern of making it difficult for adversaries to force eviction of specific entries. Consequently, this changes the code to make use of the pseudorandom iteration order of Go's maps along with the preimage resistance of the hashing function to provide the desired functionality. It has previously been discussed that the specific pseudorandom iteration order is not guaranteed by the Go spec even though in practice that is how it is implemented. This is not a concern however because even if the specific compiler doesn't implement that, the preimage resistance of the hashing function alone is enough. The following is a before and after comparison of the function for both speed and memory allocations: benchmark old ns/op new ns/op delta ---------------------------------------------------------------- BenchmarkLimitNumOrphans 3727 243 -93.48% benchmark old allocs new allocs delta ----------------------------------------------------------------- BenchmarkLimitNumOrphans 4 0 -100.00%
2016-10-24 04:14:09 +02:00
break
}
return nil
}
// addOrphan adds an orphan transaction to the orphan pool.
//
// This function MUST be called with the mempool lock held (for writes).
func (mp *TxPool) addOrphan(tx *btcutil.Tx, tag Tag) {
mempool: Optimize orphan map limiting. This optimizes the way in which the mempool oprhan map is limited in the same way the server block manager maps were previously optimized. Previously the code would read a cryptographically random value large enough to construct a hash, find the first entry larger than that value, and evict it. That approach is quite inefficient and could easily become a bottleneck when processing transactions due to the need to read from a source such as /dev/urandom and all of the subsequent hash comparisons. Luckily, strong cryptographic randomness is not needed here. The primary intent of limiting the maps is to control memory usage with a secondary concern of making it difficult for adversaries to force eviction of specific entries. Consequently, this changes the code to make use of the pseudorandom iteration order of Go's maps along with the preimage resistance of the hashing function to provide the desired functionality. It has previously been discussed that the specific pseudorandom iteration order is not guaranteed by the Go spec even though in practice that is how it is implemented. This is not a concern however because even if the specific compiler doesn't implement that, the preimage resistance of the hashing function alone is enough. The following is a before and after comparison of the function for both speed and memory allocations: benchmark old ns/op new ns/op delta ---------------------------------------------------------------- BenchmarkLimitNumOrphans 3727 243 -93.48% benchmark old allocs new allocs delta ----------------------------------------------------------------- BenchmarkLimitNumOrphans 4 0 -100.00%
2016-10-24 04:14:09 +02:00
// Nothing to do if no orphans are allowed.
if mp.cfg.Policy.MaxOrphanTxs <= 0 {
return
}
// Limit the number orphan transactions to prevent memory exhaustion.
// This will periodically remove any expired orphans and evict a random
// orphan if space is still needed.
mp.limitNumOrphans()
mp.orphans[*tx.Hash()] = &orphanTx{
tx: tx,
tag: tag,
expiration: time.Now().Add(orphanTTL),
}
for _, txIn := range tx.MsgTx().TxIn {
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
if _, exists := mp.orphansByPrev[txIn.PreviousOutPoint]; !exists {
mp.orphansByPrev[txIn.PreviousOutPoint] =
make(map[chainhash.Hash]*btcutil.Tx)
}
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
mp.orphansByPrev[txIn.PreviousOutPoint][*tx.Hash()] = tx
}
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
log.Debugf("Stored orphan transaction %v (total: %d)", tx.Hash(),
len(mp.orphans))
}
// maybeAddOrphan potentially adds an orphan to the orphan pool.
//
// This function MUST be called with the mempool lock held (for writes).
func (mp *TxPool) maybeAddOrphan(tx *btcutil.Tx, tag Tag) error {
// Ignore orphan transactions that are too large. This helps avoid
// a memory exhaustion attack based on sending a lot of really large
// orphans. In the case there is a valid transaction larger than this,
// it will ultimtely be rebroadcast after the parent transactions
// have been mined or otherwise received.
//
// Note that the number of orphan transactions in the orphan pool is
// also limited, so this equates to a maximum memory used of
// mp.cfg.Policy.MaxOrphanTxSize * mp.cfg.Policy.MaxOrphanTxs (which is ~5MB
// using the default values at the time this comment was written).
serializedLen := tx.MsgTx().SerializeSize()
if serializedLen > mp.cfg.Policy.MaxOrphanTxSize {
str := fmt.Sprintf("orphan transaction size of %d bytes is "+
"larger than max allowed size of %d bytes",
serializedLen, mp.cfg.Policy.MaxOrphanTxSize)
return txRuleError(wire.RejectNonstandard, str)
}
// Add the orphan if the none of the above disqualified it.
mp.addOrphan(tx, tag)
return nil
}
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
// removeOrphanDoubleSpends removes all orphans which spend outputs spent by the
// passed transaction from the orphan pool. Removing those orphans then leads
// to removing all orphans which rely on them, recursively. This is necessary
// when a transaction is added to the main pool because it may spend outputs
// that orphans also spend.
//
// This function MUST be called with the mempool lock held (for writes).
func (mp *TxPool) removeOrphanDoubleSpends(tx *btcutil.Tx) {
msgTx := tx.MsgTx()
for _, txIn := range msgTx.TxIn {
for _, orphan := range mp.orphansByPrev[txIn.PreviousOutPoint] {
mp.removeOrphan(orphan, true)
}
}
}
// isTransactionInPool returns whether or not the passed transaction already
// exists in the main pool.
//
// This function MUST be called with the mempool lock held (for reads).
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
func (mp *TxPool) isTransactionInPool(hash *chainhash.Hash) bool {
if _, exists := mp.pool[*hash]; exists {
return true
}
return false
}
// IsTransactionInPool returns whether or not the passed transaction already
// exists in the main pool.
//
// This function is safe for concurrent access.
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
func (mp *TxPool) IsTransactionInPool(hash *chainhash.Hash) bool {
// Protect concurrent access.
mp.mtx.RLock()
inPool := mp.isTransactionInPool(hash)
mp.mtx.RUnlock()
return inPool
}
// isOrphanInPool returns whether or not the passed transaction already exists
// in the orphan pool.
//
// This function MUST be called with the mempool lock held (for reads).
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
func (mp *TxPool) isOrphanInPool(hash *chainhash.Hash) bool {
if _, exists := mp.orphans[*hash]; exists {
return true
}
return false
}
// IsOrphanInPool returns whether or not the passed transaction already exists
// in the orphan pool.
//
// This function is safe for concurrent access.
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
func (mp *TxPool) IsOrphanInPool(hash *chainhash.Hash) bool {
// Protect concurrent access.
mp.mtx.RLock()
inPool := mp.isOrphanInPool(hash)
mp.mtx.RUnlock()
return inPool
}
// haveTransaction returns whether or not the passed transaction already exists
// in the main pool or in the orphan pool.
//
// This function MUST be called with the mempool lock held (for reads).
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
func (mp *TxPool) haveTransaction(hash *chainhash.Hash) bool {
return mp.isTransactionInPool(hash) || mp.isOrphanInPool(hash)
}
// HaveTransaction returns whether or not the passed transaction already exists
// in the main pool or in the orphan pool.
//
// This function is safe for concurrent access.
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
func (mp *TxPool) HaveTransaction(hash *chainhash.Hash) bool {
// Protect concurrent access.
mp.mtx.RLock()
haveTx := mp.haveTransaction(hash)
mp.mtx.RUnlock()
return haveTx
}
// removeTransaction is the internal function which implements the public
// RemoveTransaction. See the comment for RemoveTransaction for more details.
//
// This function MUST be called with the mempool lock held (for writes).
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
func (mp *TxPool) removeTransaction(tx *btcutil.Tx, removeRedeemers bool) {
txHash := tx.Hash()
if removeRedeemers {
// Remove any transactions which rely on this one.
for i := uint32(0); i < uint32(len(tx.MsgTx().TxOut)); i++ {
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
prevOut := wire.OutPoint{Hash: *txHash, Index: i}
if txRedeemer, exists := mp.outpoints[prevOut]; exists {
mp.removeTransaction(txRedeemer, true)
}
}
}
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
// Remove the transaction if needed.
if txDesc, exists := mp.pool[*txHash]; exists {
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
// Remove unconfirmed address index entries associated with the
// transaction if enabled.
if mp.cfg.AddrIndex != nil {
mp.cfg.AddrIndex.RemoveUnconfirmedTx(txHash)
}
// Mark the referenced outpoints as unspent by the pool.
for _, txIn := range txDesc.Tx.MsgTx().TxIn {
delete(mp.outpoints, txIn.PreviousOutPoint)
}
delete(mp.pool, *txHash)
2016-02-12 19:44:56 +01:00
atomic.StoreInt64(&mp.lastUpdated, time.Now().Unix())
}
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// RemoveTransaction removes the passed transaction from the mempool. When the
// removeRedeemers flag is set, any transactions that redeem outputs from the
// removed transaction will also be removed recursively from the mempool, as
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// they would otherwise become orphans.
//
// This function is safe for concurrent access.
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
func (mp *TxPool) RemoveTransaction(tx *btcutil.Tx, removeRedeemers bool) {
// Protect concurrent access.
mp.mtx.Lock()
mp.removeTransaction(tx, removeRedeemers)
mp.mtx.Unlock()
}
// RemoveDoubleSpends removes all transactions which spend outputs spent by the
// passed transaction from the memory pool. Removing those transactions then
// leads to removing all transactions which rely on them, recursively. This is
// necessary when a block is connected to the main chain because the block may
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
// contain transactions which were previously unknown to the memory pool.
//
// This function is safe for concurrent access.
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
func (mp *TxPool) RemoveDoubleSpends(tx *btcutil.Tx) {
// Protect concurrent access.
mp.mtx.Lock()
for _, txIn := range tx.MsgTx().TxIn {
if txRedeemer, ok := mp.outpoints[txIn.PreviousOutPoint]; ok {
if !txRedeemer.Hash().IsEqual(tx.Hash()) {
mp.removeTransaction(txRedeemer, true)
}
}
}
mp.mtx.Unlock()
}
// addTransaction adds the passed transaction to the memory pool. It should
// not be called directly as it doesn't perform any validation. This is a
// helper for maybeAcceptTransaction.
//
// This function MUST be called with the mempool lock held (for writes).
func (mp *TxPool) addTransaction(utxoView *blockchain.UtxoViewpoint, tx *btcutil.Tx, height int32, fee int64) *TxDesc {
// Add the transaction to the pool and mark the referenced outpoints
// as spent by the pool.
txD := &TxDesc{
TxDesc: mining.TxDesc{
Tx: tx,
Added: time.Now(),
Height: height,
Fee: fee,
FeePerKB: fee * 1000 / GetTxVirtualSize(tx),
},
StartingPriority: mining.CalcPriority(tx.MsgTx(), utxoView, height),
}
mp.pool[*tx.Hash()] = txD
for _, txIn := range tx.MsgTx().TxIn {
mp.outpoints[txIn.PreviousOutPoint] = tx
}
2016-02-12 19:44:56 +01:00
atomic.StoreInt64(&mp.lastUpdated, time.Now().Unix())
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
// Add unconfirmed address index entries associated with the transaction
// if enabled.
if mp.cfg.AddrIndex != nil {
mp.cfg.AddrIndex.AddUnconfirmedTx(tx, utxoView)
}
// Record this tx for fee estimation if enabled.
if mp.cfg.FeeEstimator != nil {
mp.cfg.FeeEstimator.ObserveTransaction(txD)
}
return txD
}
// checkPoolDoubleSpend checks whether or not the passed transaction is
// attempting to spend coins already spent by other transactions in the pool.
// If it does, we'll check whether each of those transactions are signaling for
// replacement. If just one of them isn't, an error is returned. Otherwise, a
// boolean is returned signaling that the transaction is a replacement. Note it
// does not check for double spends against transactions already in the main
// chain.
//
// This function MUST be called with the mempool lock held (for reads).
func (mp *TxPool) checkPoolDoubleSpend(tx *btcutil.Tx) (bool, error) {
var isReplacement bool
for _, txIn := range tx.MsgTx().TxIn {
conflict, ok := mp.outpoints[txIn.PreviousOutPoint]
if !ok {
continue
}
// Reject the transaction if we don't accept replacement
// transactions or if it doesn't signal replacement.
if mp.cfg.Policy.RejectReplacement ||
!mp.signalsReplacement(conflict, nil) {
str := fmt.Sprintf("output %v already spent by "+
"transaction %v in the memory pool",
txIn.PreviousOutPoint, conflict.Hash())
return false, txRuleError(wire.RejectDuplicate, str)
}
isReplacement = true
}
return isReplacement, nil
}
// signalsReplacement determines if a transaction is signaling that it can be
// replaced using the Replace-By-Fee (RBF) policy. This policy specifies two
// ways a transaction can signal that it is replaceable:
//
// Explicit signaling: A transaction is considered to have opted in to allowing
// replacement of itself if any of its inputs have a sequence number less than
// 0xfffffffe.
//
// Inherited signaling: Transactions that don't explicitly signal replaceability
// are replaceable under this policy for as long as any one of their ancestors
// signals replaceability and remains unconfirmed.
//
// The cache is optional and serves as an optimization to avoid visiting
// transactions we've already determined don't signal replacement.
//
// This function MUST be called with the mempool lock held (for reads).
func (mp *TxPool) signalsReplacement(tx *btcutil.Tx,
cache map[chainhash.Hash]struct{}) bool {
// If a cache was not provided, we'll initialize one now to use for the
// recursive calls.
if cache == nil {
cache = make(map[chainhash.Hash]struct{})
}
for _, txIn := range tx.MsgTx().TxIn {
if txIn.Sequence <= MaxRBFSequence {
return true
}
hash := txIn.PreviousOutPoint.Hash
unconfirmedAncestor, ok := mp.pool[hash]
if !ok {
continue
}
// If we've already determined the transaction doesn't signal
// replacement, we can avoid visiting it again.
if _, ok := cache[hash]; ok {
continue
}
if mp.signalsReplacement(unconfirmedAncestor.Tx, cache) {
return true
}
// Since the transaction doesn't signal replacement, we'll cache
// its result to ensure we don't attempt to determine so again.
cache[hash] = struct{}{}
}
return false
}
// txAncestors returns all of the unconfirmed ancestors of the given
// transaction. Given transactions A, B, and C where C spends B and B spends A,
// A and B are considered ancestors of C.
//
// The cache is optional and serves as an optimization to avoid visiting
// transactions we've already determined ancestors of.
//
// This function MUST be called with the mempool lock held (for reads).
func (mp *TxPool) txAncestors(tx *btcutil.Tx,
cache map[chainhash.Hash]map[chainhash.Hash]*btcutil.Tx) map[chainhash.Hash]*btcutil.Tx {
// If a cache was not provided, we'll initialize one now to use for the
// recursive calls.
if cache == nil {
cache = make(map[chainhash.Hash]map[chainhash.Hash]*btcutil.Tx)
}
ancestors := make(map[chainhash.Hash]*btcutil.Tx)
for _, txIn := range tx.MsgTx().TxIn {
parent, ok := mp.pool[txIn.PreviousOutPoint.Hash]
if !ok {
continue
}
ancestors[*parent.Tx.Hash()] = parent.Tx
// Determine if the ancestors of this ancestor have already been
// computed. If they haven't, we'll do so now and cache them to
// use them later on if necessary.
moreAncestors, ok := cache[*parent.Tx.Hash()]
if !ok {
moreAncestors = mp.txAncestors(parent.Tx, cache)
cache[*parent.Tx.Hash()] = moreAncestors
}
for hash, ancestor := range moreAncestors {
ancestors[hash] = ancestor
}
}
return ancestors
}
// txDescendants returns all of the unconfirmed descendants of the given
// transaction. Given transactions A, B, and C where C spends B and B spends A,
// B and C are considered descendants of A. A cache can be provided in order to
// easily retrieve the descendants of transactions we've already determined the
// descendants of.
//
// This function MUST be called with the mempool lock held (for reads).
func (mp *TxPool) txDescendants(tx *btcutil.Tx,
cache map[chainhash.Hash]map[chainhash.Hash]*btcutil.Tx) map[chainhash.Hash]*btcutil.Tx {
// If a cache was not provided, we'll initialize one now to use for the
// recursive calls.
if cache == nil {
cache = make(map[chainhash.Hash]map[chainhash.Hash]*btcutil.Tx)
}
// We'll go through all of the outputs of the transaction to determine
// if they are spent by any other mempool transactions.
descendants := make(map[chainhash.Hash]*btcutil.Tx)
op := wire.OutPoint{Hash: *tx.Hash()}
for i := range tx.MsgTx().TxOut {
op.Index = uint32(i)
descendant, ok := mp.outpoints[op]
if !ok {
continue
}
descendants[*descendant.Hash()] = descendant
// Determine if the descendants of this descendant have already
// been computed. If they haven't, we'll do so now and cache
// them to use them later on if necessary.
moreDescendants, ok := cache[*descendant.Hash()]
if !ok {
moreDescendants = mp.txDescendants(descendant, cache)
cache[*descendant.Hash()] = moreDescendants
}
for _, moreDescendant := range moreDescendants {
descendants[*moreDescendant.Hash()] = moreDescendant
}
}
return descendants
}
// txConflicts returns all of the unconfirmed transactions that would become
// conflicts if we were to accept the given transaction into the mempool. An
// unconfirmed conflict is known as a transaction that spends an output already
// spent by a different transaction within the mempool. Any descendants of these
// transactions are also considered conflicts as they would no longer exist.
// These are generally not allowed except for transactions that signal RBF
// support.
//
// This function MUST be called with the mempool lock held (for reads).
func (mp *TxPool) txConflicts(tx *btcutil.Tx) map[chainhash.Hash]*btcutil.Tx {
conflicts := make(map[chainhash.Hash]*btcutil.Tx)
for _, txIn := range tx.MsgTx().TxIn {
conflict, ok := mp.outpoints[txIn.PreviousOutPoint]
if !ok {
continue
}
conflicts[*conflict.Hash()] = conflict
for hash, descendant := range mp.txDescendants(conflict, nil) {
conflicts[hash] = descendant
}
}
return conflicts
}
2018-05-19 03:55:52 +02:00
// CheckSpend checks whether the passed outpoint is already spent by a
// transaction in the mempool. If that's the case the spending transaction will
// be returned, if not nil will be returned.
func (mp *TxPool) CheckSpend(op wire.OutPoint) *btcutil.Tx {
mp.mtx.RLock()
txR := mp.outpoints[op]
mp.mtx.RUnlock()
return txR
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// fetchInputUtxos loads utxo details about the input transactions referenced by
// the passed transaction. First, it loads the details form the viewpoint of
// the main chain, then it adjusts them based upon the contents of the
// transaction pool.
//
// This function MUST be called with the mempool lock held (for reads).
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
func (mp *TxPool) fetchInputUtxos(tx *btcutil.Tx) (*blockchain.UtxoViewpoint, error) {
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
utxoView, err := mp.cfg.FetchUtxoView(tx)
if err != nil {
return nil, err
}
// Attempt to populate any missing inputs from the transaction pool.
multi: Rework utxoset/view to use outpoints. This modifies the utxoset in the database and related UtxoViewpoint to store and work with unspent transaction outputs on a per-output basis instead of at a transaction level. This was inspired by similar recent changes in Bitcoin Core. The primary motivation is to simplify the code, pave the way for a utxo cache, and generally focus on optimizing runtime performance. The tradeoff is that this approach does somewhat increase the size of the serialized utxoset since it means that the transaction hash is duplicated for each output as a part of the key and some additional details such as whether the containing transaction is a coinbase and the block height it was a part of are duplicated in each output. However, in practice, the size difference isn't all that large, disk space is relatively cheap, certainly cheaper than memory, and it is much more important to provide more efficient runtime operation since that is the ultimate purpose of the daemon. While performing this conversion, it also simplifies the code to remove the transaction version information from the utxoset as well as the spend journal. The logic for only serializing it under certain circumstances is complicated and it isn't actually used anywhere aside from the gettxout RPC where it also isn't used by anything important either. Consequently, this also removes the version field of the gettxout RPC result. The utxos in the database are automatically migrated to the new format with this commit and it is possible to interrupt and resume the migration process. Finally, it also updates the tests for the new format and adds a new function to the tests to convert the old test data to the new format for convenience. The data has already been converted and updated in the commit. An overview of the changes are as follows: - Remove transaction version from both spent and unspent output entries - Update utxo serialization format to exclude the version - Modify the spend journal serialization format - The old version field is now reserved and always stores zero and ignores it when reading - This allows old entries to be used by new code without having to migrate the entire spend journal - Remove version field from gettxout RPC result - Convert UtxoEntry to represent a specific utxo instead of a transaction with all remaining utxos - Optimize for memory usage with an eye towards a utxo cache - Combine details such as whether the txout was contained in a coinbase, is spent, and is modified into a single packed field of bit flags - Align entry fields to eliminate extra padding since ultimately there will be a lot of these in memory - Introduce a free list for serializing an outpoint to the database key format to significantly reduce pressure on the GC - Update all related functions that previously dealt with transaction hashes to accept outpoints instead - Update all callers accordingly - Only add individually requested outputs from the mempool when constructing a mempool view - Modify the spend journal to always store the block height and coinbase information with every spent txout - Introduce code to handle fetching the missing information from another utxo from the same transaction in the event an old style entry is encountered - Make use of a database cursor with seek to do this much more efficiently than testing every possible output - Always decompress data loaded from the database now that a utxo entry only consists of a specific output - Introduce upgrade code to migrate the utxo set to the new format - Store versions of the utxoset and spend journal buckets - Allow migration process to be interrupted and resumed - Update all tests to expect the correct encodings, remove tests that no longer apply, and add new ones for the new expected behavior - Convert old tests for the legacy utxo format deserialization code to test the new function that is used during upgrade - Update the utxostore test data and add function that was used to convert it - Introduce a few new functions on UtxoViewpoint - AddTxOut for adding an individual txout versus all of them - addTxOut to handle the common code between the new AddTxOut and existing AddTxOuts - RemoveEntry for removing an individual txout - fetchEntryByHash for fetching any remaining utxo for a given transaction hash
2017-09-03 09:59:15 +02:00
for _, txIn := range tx.MsgTx().TxIn {
prevOut := &txIn.PreviousOutPoint
entry := utxoView.LookupEntry(*prevOut)
if entry != nil && !entry.IsSpent() {
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
continue
}
multi: Rework utxoset/view to use outpoints. This modifies the utxoset in the database and related UtxoViewpoint to store and work with unspent transaction outputs on a per-output basis instead of at a transaction level. This was inspired by similar recent changes in Bitcoin Core. The primary motivation is to simplify the code, pave the way for a utxo cache, and generally focus on optimizing runtime performance. The tradeoff is that this approach does somewhat increase the size of the serialized utxoset since it means that the transaction hash is duplicated for each output as a part of the key and some additional details such as whether the containing transaction is a coinbase and the block height it was a part of are duplicated in each output. However, in practice, the size difference isn't all that large, disk space is relatively cheap, certainly cheaper than memory, and it is much more important to provide more efficient runtime operation since that is the ultimate purpose of the daemon. While performing this conversion, it also simplifies the code to remove the transaction version information from the utxoset as well as the spend journal. The logic for only serializing it under certain circumstances is complicated and it isn't actually used anywhere aside from the gettxout RPC where it also isn't used by anything important either. Consequently, this also removes the version field of the gettxout RPC result. The utxos in the database are automatically migrated to the new format with this commit and it is possible to interrupt and resume the migration process. Finally, it also updates the tests for the new format and adds a new function to the tests to convert the old test data to the new format for convenience. The data has already been converted and updated in the commit. An overview of the changes are as follows: - Remove transaction version from both spent and unspent output entries - Update utxo serialization format to exclude the version - Modify the spend journal serialization format - The old version field is now reserved and always stores zero and ignores it when reading - This allows old entries to be used by new code without having to migrate the entire spend journal - Remove version field from gettxout RPC result - Convert UtxoEntry to represent a specific utxo instead of a transaction with all remaining utxos - Optimize for memory usage with an eye towards a utxo cache - Combine details such as whether the txout was contained in a coinbase, is spent, and is modified into a single packed field of bit flags - Align entry fields to eliminate extra padding since ultimately there will be a lot of these in memory - Introduce a free list for serializing an outpoint to the database key format to significantly reduce pressure on the GC - Update all related functions that previously dealt with transaction hashes to accept outpoints instead - Update all callers accordingly - Only add individually requested outputs from the mempool when constructing a mempool view - Modify the spend journal to always store the block height and coinbase information with every spent txout - Introduce code to handle fetching the missing information from another utxo from the same transaction in the event an old style entry is encountered - Make use of a database cursor with seek to do this much more efficiently than testing every possible output - Always decompress data loaded from the database now that a utxo entry only consists of a specific output - Introduce upgrade code to migrate the utxo set to the new format - Store versions of the utxoset and spend journal buckets - Allow migration process to be interrupted and resumed - Update all tests to expect the correct encodings, remove tests that no longer apply, and add new ones for the new expected behavior - Convert old tests for the legacy utxo format deserialization code to test the new function that is used during upgrade - Update the utxostore test data and add function that was used to convert it - Introduce a few new functions on UtxoViewpoint - AddTxOut for adding an individual txout versus all of them - addTxOut to handle the common code between the new AddTxOut and existing AddTxOuts - RemoveEntry for removing an individual txout - fetchEntryByHash for fetching any remaining utxo for a given transaction hash
2017-09-03 09:59:15 +02:00
if poolTxDesc, exists := mp.pool[prevOut.Hash]; exists {
// AddTxOut ignores out of range index values, so it is
// safe to call without bounds checking here.
utxoView.AddTxOut(poolTxDesc.Tx, prevOut.Index,
mining.UnminedHeight)
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
}
}
multi: Rework utxoset/view to use outpoints. This modifies the utxoset in the database and related UtxoViewpoint to store and work with unspent transaction outputs on a per-output basis instead of at a transaction level. This was inspired by similar recent changes in Bitcoin Core. The primary motivation is to simplify the code, pave the way for a utxo cache, and generally focus on optimizing runtime performance. The tradeoff is that this approach does somewhat increase the size of the serialized utxoset since it means that the transaction hash is duplicated for each output as a part of the key and some additional details such as whether the containing transaction is a coinbase and the block height it was a part of are duplicated in each output. However, in practice, the size difference isn't all that large, disk space is relatively cheap, certainly cheaper than memory, and it is much more important to provide more efficient runtime operation since that is the ultimate purpose of the daemon. While performing this conversion, it also simplifies the code to remove the transaction version information from the utxoset as well as the spend journal. The logic for only serializing it under certain circumstances is complicated and it isn't actually used anywhere aside from the gettxout RPC where it also isn't used by anything important either. Consequently, this also removes the version field of the gettxout RPC result. The utxos in the database are automatically migrated to the new format with this commit and it is possible to interrupt and resume the migration process. Finally, it also updates the tests for the new format and adds a new function to the tests to convert the old test data to the new format for convenience. The data has already been converted and updated in the commit. An overview of the changes are as follows: - Remove transaction version from both spent and unspent output entries - Update utxo serialization format to exclude the version - Modify the spend journal serialization format - The old version field is now reserved and always stores zero and ignores it when reading - This allows old entries to be used by new code without having to migrate the entire spend journal - Remove version field from gettxout RPC result - Convert UtxoEntry to represent a specific utxo instead of a transaction with all remaining utxos - Optimize for memory usage with an eye towards a utxo cache - Combine details such as whether the txout was contained in a coinbase, is spent, and is modified into a single packed field of bit flags - Align entry fields to eliminate extra padding since ultimately there will be a lot of these in memory - Introduce a free list for serializing an outpoint to the database key format to significantly reduce pressure on the GC - Update all related functions that previously dealt with transaction hashes to accept outpoints instead - Update all callers accordingly - Only add individually requested outputs from the mempool when constructing a mempool view - Modify the spend journal to always store the block height and coinbase information with every spent txout - Introduce code to handle fetching the missing information from another utxo from the same transaction in the event an old style entry is encountered - Make use of a database cursor with seek to do this much more efficiently than testing every possible output - Always decompress data loaded from the database now that a utxo entry only consists of a specific output - Introduce upgrade code to migrate the utxo set to the new format - Store versions of the utxoset and spend journal buckets - Allow migration process to be interrupted and resumed - Update all tests to expect the correct encodings, remove tests that no longer apply, and add new ones for the new expected behavior - Convert old tests for the legacy utxo format deserialization code to test the new function that is used during upgrade - Update the utxostore test data and add function that was used to convert it - Introduce a few new functions on UtxoViewpoint - AddTxOut for adding an individual txout versus all of them - addTxOut to handle the common code between the new AddTxOut and existing AddTxOuts - RemoveEntry for removing an individual txout - fetchEntryByHash for fetching any remaining utxo for a given transaction hash
2017-09-03 09:59:15 +02:00
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
return utxoView, nil
}
// FetchTransaction returns the requested transaction from the transaction pool.
// This only fetches from the main transaction pool and does not include
// orphans.
//
// This function is safe for concurrent access.
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
func (mp *TxPool) FetchTransaction(txHash *chainhash.Hash) (*btcutil.Tx, error) {
// Protect concurrent access.
mp.mtx.RLock()
txDesc, exists := mp.pool[*txHash]
mp.mtx.RUnlock()
if exists {
return txDesc.Tx, nil
}
return nil, fmt.Errorf("transaction is not in the pool")
}
// validateReplacement determines whether a transaction is deemed as a valid
// replacement of all of its conflicts according to the RBF policy. If it is
// valid, no error is returned. Otherwise, an error is returned indicating what
// went wrong.
//
// This function MUST be called with the mempool lock held (for reads).
func (mp *TxPool) validateReplacement(tx *btcutil.Tx,
txFee int64) (map[chainhash.Hash]*btcutil.Tx, error) {
// First, we'll make sure the set of conflicting transactions doesn't
// exceed the maximum allowed.
conflicts := mp.txConflicts(tx)
if len(conflicts) > MaxReplacementEvictions {
str := fmt.Sprintf("replacement transaction %v evicts more "+
"transactions than permitted: max is %v, evicts %v",
tx.Hash(), MaxReplacementEvictions, len(conflicts))
return nil, txRuleError(wire.RejectNonstandard, str)
}
// The set of conflicts (transactions we'll replace) and ancestors
// should not overlap, otherwise the replacement would be spending an
// output that no longer exists.
for ancestorHash := range mp.txAncestors(tx, nil) {
if _, ok := conflicts[ancestorHash]; !ok {
continue
}
str := fmt.Sprintf("replacement transaction %v spends parent "+
"transaction %v", tx.Hash(), ancestorHash)
return nil, txRuleError(wire.RejectInvalid, str)
}
// The replacement should have a higher fee rate than each of the
// conflicting transactions and a higher absolute fee than the fee sum
// of all the conflicting transactions.
//
// We usually don't want to accept replacements with lower fee rates
// than what they replaced as that would lower the fee rate of the next
// block. Requiring that the fee rate always be increased is also an
// easy-to-reason about way to prevent DoS attacks via replacements.
var (
txSize = GetTxVirtualSize(tx)
txFeeRate = txFee * 1000 / txSize
conflictsFee int64
conflictsParents = make(map[chainhash.Hash]struct{})
)
for hash, conflict := range conflicts {
if txFeeRate <= mp.pool[hash].FeePerKB {
str := fmt.Sprintf("replacement transaction %v has an "+
"insufficient fee rate: needs more than %v, "+
"has %v", tx.Hash(), mp.pool[hash].FeePerKB,
txFeeRate)
return nil, txRuleError(wire.RejectInsufficientFee, str)
}
conflictsFee += mp.pool[hash].Fee
// We'll track each conflict's parents to ensure the replacement
// isn't spending any new unconfirmed inputs.
for _, txIn := range conflict.MsgTx().TxIn {
conflictsParents[txIn.PreviousOutPoint.Hash] = struct{}{}
}
}
// It should also have an absolute fee greater than all of the
// transactions it intends to replace and pay for its own bandwidth,
// which is determined by our minimum relay fee.
minFee := calcMinRequiredTxRelayFee(txSize, mp.cfg.Policy.MinRelayTxFee)
if txFee < conflictsFee+minFee {
str := fmt.Sprintf("replacement transaction %v has an "+
"insufficient absolute fee: needs %v, has %v",
tx.Hash(), conflictsFee+minFee, txFee)
return nil, txRuleError(wire.RejectInsufficientFee, str)
}
// Finally, it should not spend any new unconfirmed outputs, other than
// the ones already included in the parents of the conflicting
// transactions it'll replace.
for _, txIn := range tx.MsgTx().TxIn {
if _, ok := conflictsParents[txIn.PreviousOutPoint.Hash]; ok {
continue
}
// Confirmed outputs are valid to spend in the replacement.
if _, ok := mp.pool[txIn.PreviousOutPoint.Hash]; !ok {
continue
}
str := fmt.Sprintf("replacement transaction spends new "+
"unconfirmed input %v not found in conflicting "+
"transactions", txIn.PreviousOutPoint)
return nil, txRuleError(wire.RejectInvalid, str)
}
return conflicts, nil
}
// maybeAcceptTransaction is the internal function which implements the public
// MaybeAcceptTransaction. See the comment for MaybeAcceptTransaction for
// more details.
//
// This function MUST be called with the mempool lock held (for writes).
func (mp *TxPool) maybeAcceptTransaction(tx *btcutil.Tx, isNew, rateLimit, rejectDupOrphans bool) ([]*chainhash.Hash, *TxDesc, error) {
txHash := tx.Hash()
// If a transaction has iwtness data, and segwit isn't active yet, If
// segwit isn't active yet, then we won't accept it into the mempool as
// it can't be mined yet.
if tx.MsgTx().HasWitness() {
segwitActive, err := mp.cfg.IsDeploymentActive(chaincfg.DeploymentSegwit)
if err != nil {
return nil, nil, err
}
if !segwitActive {
str := fmt.Sprintf("transaction %v has witness data, "+
"but segwit isn't active yet", txHash)
return nil, nil, txRuleError(wire.RejectNonstandard, str)
}
}
// Don't accept the transaction if it already exists in the pool. This
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
// applies to orphan transactions as well when the reject duplicate
// orphans flag is set. This check is intended to be a quick check to
// weed out duplicates.
if mp.isTransactionInPool(txHash) || (rejectDupOrphans &&
mp.isOrphanInPool(txHash)) {
str := fmt.Sprintf("already have transaction %v", txHash)
return nil, nil, txRuleError(wire.RejectDuplicate, str)
}
// Perform preliminary sanity checks on the transaction. This makes
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
// use of blockchain which contains the invariant rules for what
// transactions are allowed into blocks.
err := blockchain.CheckTransactionSanity(tx)
if err != nil {
if cerr, ok := err.(blockchain.RuleError); ok {
return nil, nil, chainRuleError(cerr)
}
return nil, nil, err
}
// A standalone transaction must not be a coinbase transaction.
if blockchain.IsCoinBase(tx) {
str := fmt.Sprintf("transaction %v is an individual coinbase",
txHash)
return nil, nil, txRuleError(wire.RejectInvalid, str)
}
// Get the current height of the main chain. A standalone transaction
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// will be mined into the next block at best, so its height is at least
// one more than the current height.
bestHeight := mp.cfg.BestHeight()
nextBlockHeight := bestHeight + 1
medianTimePast := mp.cfg.MedianTimePast()
// Don't allow non-standard transactions if the network parameters
// forbid their acceptance.
if !mp.cfg.Policy.AcceptNonStd {
err = checkTransactionStandard(tx, nextBlockHeight,
medianTimePast, mp.cfg.Policy.MinRelayTxFee,
mp.cfg.Policy.MaxTxVersion)
if err != nil {
// Attempt to extract a reject code from the error so
// it can be retained. When not possible, fall back to
// a non standard error.
rejectCode, found := extractRejectCode(err)
if !found {
rejectCode = wire.RejectNonstandard
}
str := fmt.Sprintf("transaction %v is not standard: %v",
txHash, err)
return nil, nil, txRuleError(rejectCode, str)
}
}
// The transaction may not use any of the same outputs as other
// transactions already in the pool as that would ultimately result in a
// double spend, unless those transactions signal for RBF. This check is
// intended to be quick and therefore only detects double spends within
// the transaction pool itself. The transaction could still be double
// spending coins from the main chain at this point. There is a more
// in-depth check that happens later after fetching the referenced
// transaction inputs from the main chain which examines the actual
// spend data and prevents double spends.
isReplacement, err := mp.checkPoolDoubleSpend(tx)
if err != nil {
return nil, nil, err
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// Fetch all of the unspent transaction outputs referenced by the inputs
// to this transaction. This function also attempts to fetch the
// transaction itself to be used for detecting a duplicate transaction
// without needing to do a separate lookup.
utxoView, err := mp.fetchInputUtxos(tx)
if err != nil {
if cerr, ok := err.(blockchain.RuleError); ok {
return nil, nil, chainRuleError(cerr)
}
return nil, nil, err
}
// Don't allow the transaction if it exists in the main chain and is not
// not already fully spent.
multi: Rework utxoset/view to use outpoints. This modifies the utxoset in the database and related UtxoViewpoint to store and work with unspent transaction outputs on a per-output basis instead of at a transaction level. This was inspired by similar recent changes in Bitcoin Core. The primary motivation is to simplify the code, pave the way for a utxo cache, and generally focus on optimizing runtime performance. The tradeoff is that this approach does somewhat increase the size of the serialized utxoset since it means that the transaction hash is duplicated for each output as a part of the key and some additional details such as whether the containing transaction is a coinbase and the block height it was a part of are duplicated in each output. However, in practice, the size difference isn't all that large, disk space is relatively cheap, certainly cheaper than memory, and it is much more important to provide more efficient runtime operation since that is the ultimate purpose of the daemon. While performing this conversion, it also simplifies the code to remove the transaction version information from the utxoset as well as the spend journal. The logic for only serializing it under certain circumstances is complicated and it isn't actually used anywhere aside from the gettxout RPC where it also isn't used by anything important either. Consequently, this also removes the version field of the gettxout RPC result. The utxos in the database are automatically migrated to the new format with this commit and it is possible to interrupt and resume the migration process. Finally, it also updates the tests for the new format and adds a new function to the tests to convert the old test data to the new format for convenience. The data has already been converted and updated in the commit. An overview of the changes are as follows: - Remove transaction version from both spent and unspent output entries - Update utxo serialization format to exclude the version - Modify the spend journal serialization format - The old version field is now reserved and always stores zero and ignores it when reading - This allows old entries to be used by new code without having to migrate the entire spend journal - Remove version field from gettxout RPC result - Convert UtxoEntry to represent a specific utxo instead of a transaction with all remaining utxos - Optimize for memory usage with an eye towards a utxo cache - Combine details such as whether the txout was contained in a coinbase, is spent, and is modified into a single packed field of bit flags - Align entry fields to eliminate extra padding since ultimately there will be a lot of these in memory - Introduce a free list for serializing an outpoint to the database key format to significantly reduce pressure on the GC - Update all related functions that previously dealt with transaction hashes to accept outpoints instead - Update all callers accordingly - Only add individually requested outputs from the mempool when constructing a mempool view - Modify the spend journal to always store the block height and coinbase information with every spent txout - Introduce code to handle fetching the missing information from another utxo from the same transaction in the event an old style entry is encountered - Make use of a database cursor with seek to do this much more efficiently than testing every possible output - Always decompress data loaded from the database now that a utxo entry only consists of a specific output - Introduce upgrade code to migrate the utxo set to the new format - Store versions of the utxoset and spend journal buckets - Allow migration process to be interrupted and resumed - Update all tests to expect the correct encodings, remove tests that no longer apply, and add new ones for the new expected behavior - Convert old tests for the legacy utxo format deserialization code to test the new function that is used during upgrade - Update the utxostore test data and add function that was used to convert it - Introduce a few new functions on UtxoViewpoint - AddTxOut for adding an individual txout versus all of them - addTxOut to handle the common code between the new AddTxOut and existing AddTxOuts - RemoveEntry for removing an individual txout - fetchEntryByHash for fetching any remaining utxo for a given transaction hash
2017-09-03 09:59:15 +02:00
prevOut := wire.OutPoint{Hash: *txHash}
for txOutIdx := range tx.MsgTx().TxOut {
prevOut.Index = uint32(txOutIdx)
entry := utxoView.LookupEntry(prevOut)
if entry != nil && !entry.IsSpent() {
return nil, nil, txRuleError(wire.RejectDuplicate,
"transaction already exists")
}
utxoView.RemoveEntry(prevOut)
}
multi: Rework utxoset/view to use outpoints. This modifies the utxoset in the database and related UtxoViewpoint to store and work with unspent transaction outputs on a per-output basis instead of at a transaction level. This was inspired by similar recent changes in Bitcoin Core. The primary motivation is to simplify the code, pave the way for a utxo cache, and generally focus on optimizing runtime performance. The tradeoff is that this approach does somewhat increase the size of the serialized utxoset since it means that the transaction hash is duplicated for each output as a part of the key and some additional details such as whether the containing transaction is a coinbase and the block height it was a part of are duplicated in each output. However, in practice, the size difference isn't all that large, disk space is relatively cheap, certainly cheaper than memory, and it is much more important to provide more efficient runtime operation since that is the ultimate purpose of the daemon. While performing this conversion, it also simplifies the code to remove the transaction version information from the utxoset as well as the spend journal. The logic for only serializing it under certain circumstances is complicated and it isn't actually used anywhere aside from the gettxout RPC where it also isn't used by anything important either. Consequently, this also removes the version field of the gettxout RPC result. The utxos in the database are automatically migrated to the new format with this commit and it is possible to interrupt and resume the migration process. Finally, it also updates the tests for the new format and adds a new function to the tests to convert the old test data to the new format for convenience. The data has already been converted and updated in the commit. An overview of the changes are as follows: - Remove transaction version from both spent and unspent output entries - Update utxo serialization format to exclude the version - Modify the spend journal serialization format - The old version field is now reserved and always stores zero and ignores it when reading - This allows old entries to be used by new code without having to migrate the entire spend journal - Remove version field from gettxout RPC result - Convert UtxoEntry to represent a specific utxo instead of a transaction with all remaining utxos - Optimize for memory usage with an eye towards a utxo cache - Combine details such as whether the txout was contained in a coinbase, is spent, and is modified into a single packed field of bit flags - Align entry fields to eliminate extra padding since ultimately there will be a lot of these in memory - Introduce a free list for serializing an outpoint to the database key format to significantly reduce pressure on the GC - Update all related functions that previously dealt with transaction hashes to accept outpoints instead - Update all callers accordingly - Only add individually requested outputs from the mempool when constructing a mempool view - Modify the spend journal to always store the block height and coinbase information with every spent txout - Introduce code to handle fetching the missing information from another utxo from the same transaction in the event an old style entry is encountered - Make use of a database cursor with seek to do this much more efficiently than testing every possible output - Always decompress data loaded from the database now that a utxo entry only consists of a specific output - Introduce upgrade code to migrate the utxo set to the new format - Store versions of the utxoset and spend journal buckets - Allow migration process to be interrupted and resumed - Update all tests to expect the correct encodings, remove tests that no longer apply, and add new ones for the new expected behavior - Convert old tests for the legacy utxo format deserialization code to test the new function that is used during upgrade - Update the utxostore test data and add function that was used to convert it - Introduce a few new functions on UtxoViewpoint - AddTxOut for adding an individual txout versus all of them - addTxOut to handle the common code between the new AddTxOut and existing AddTxOuts - RemoveEntry for removing an individual txout - fetchEntryByHash for fetching any remaining utxo for a given transaction hash
2017-09-03 09:59:15 +02:00
// Transaction is an orphan if any of the referenced transaction outputs
// don't exist or are already spent. Adding orphans to the orphan pool
// is not handled by this function, and the caller should use
// maybeAddOrphan if this behavior is desired.
var missingParents []*chainhash.Hash
multi: Rework utxoset/view to use outpoints. This modifies the utxoset in the database and related UtxoViewpoint to store and work with unspent transaction outputs on a per-output basis instead of at a transaction level. This was inspired by similar recent changes in Bitcoin Core. The primary motivation is to simplify the code, pave the way for a utxo cache, and generally focus on optimizing runtime performance. The tradeoff is that this approach does somewhat increase the size of the serialized utxoset since it means that the transaction hash is duplicated for each output as a part of the key and some additional details such as whether the containing transaction is a coinbase and the block height it was a part of are duplicated in each output. However, in practice, the size difference isn't all that large, disk space is relatively cheap, certainly cheaper than memory, and it is much more important to provide more efficient runtime operation since that is the ultimate purpose of the daemon. While performing this conversion, it also simplifies the code to remove the transaction version information from the utxoset as well as the spend journal. The logic for only serializing it under certain circumstances is complicated and it isn't actually used anywhere aside from the gettxout RPC where it also isn't used by anything important either. Consequently, this also removes the version field of the gettxout RPC result. The utxos in the database are automatically migrated to the new format with this commit and it is possible to interrupt and resume the migration process. Finally, it also updates the tests for the new format and adds a new function to the tests to convert the old test data to the new format for convenience. The data has already been converted and updated in the commit. An overview of the changes are as follows: - Remove transaction version from both spent and unspent output entries - Update utxo serialization format to exclude the version - Modify the spend journal serialization format - The old version field is now reserved and always stores zero and ignores it when reading - This allows old entries to be used by new code without having to migrate the entire spend journal - Remove version field from gettxout RPC result - Convert UtxoEntry to represent a specific utxo instead of a transaction with all remaining utxos - Optimize for memory usage with an eye towards a utxo cache - Combine details such as whether the txout was contained in a coinbase, is spent, and is modified into a single packed field of bit flags - Align entry fields to eliminate extra padding since ultimately there will be a lot of these in memory - Introduce a free list for serializing an outpoint to the database key format to significantly reduce pressure on the GC - Update all related functions that previously dealt with transaction hashes to accept outpoints instead - Update all callers accordingly - Only add individually requested outputs from the mempool when constructing a mempool view - Modify the spend journal to always store the block height and coinbase information with every spent txout - Introduce code to handle fetching the missing information from another utxo from the same transaction in the event an old style entry is encountered - Make use of a database cursor with seek to do this much more efficiently than testing every possible output - Always decompress data loaded from the database now that a utxo entry only consists of a specific output - Introduce upgrade code to migrate the utxo set to the new format - Store versions of the utxoset and spend journal buckets - Allow migration process to be interrupted and resumed - Update all tests to expect the correct encodings, remove tests that no longer apply, and add new ones for the new expected behavior - Convert old tests for the legacy utxo format deserialization code to test the new function that is used during upgrade - Update the utxostore test data and add function that was used to convert it - Introduce a few new functions on UtxoViewpoint - AddTxOut for adding an individual txout versus all of them - addTxOut to handle the common code between the new AddTxOut and existing AddTxOuts - RemoveEntry for removing an individual txout - fetchEntryByHash for fetching any remaining utxo for a given transaction hash
2017-09-03 09:59:15 +02:00
for outpoint, entry := range utxoView.Entries() {
if entry == nil || entry.IsSpent() {
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// Must make a copy of the hash here since the iterator
// is replaced and taking its address directly would
// result in all of the entries pointing to the same
// memory location and thus all be the final hash.
multi: Rework utxoset/view to use outpoints. This modifies the utxoset in the database and related UtxoViewpoint to store and work with unspent transaction outputs on a per-output basis instead of at a transaction level. This was inspired by similar recent changes in Bitcoin Core. The primary motivation is to simplify the code, pave the way for a utxo cache, and generally focus on optimizing runtime performance. The tradeoff is that this approach does somewhat increase the size of the serialized utxoset since it means that the transaction hash is duplicated for each output as a part of the key and some additional details such as whether the containing transaction is a coinbase and the block height it was a part of are duplicated in each output. However, in practice, the size difference isn't all that large, disk space is relatively cheap, certainly cheaper than memory, and it is much more important to provide more efficient runtime operation since that is the ultimate purpose of the daemon. While performing this conversion, it also simplifies the code to remove the transaction version information from the utxoset as well as the spend journal. The logic for only serializing it under certain circumstances is complicated and it isn't actually used anywhere aside from the gettxout RPC where it also isn't used by anything important either. Consequently, this also removes the version field of the gettxout RPC result. The utxos in the database are automatically migrated to the new format with this commit and it is possible to interrupt and resume the migration process. Finally, it also updates the tests for the new format and adds a new function to the tests to convert the old test data to the new format for convenience. The data has already been converted and updated in the commit. An overview of the changes are as follows: - Remove transaction version from both spent and unspent output entries - Update utxo serialization format to exclude the version - Modify the spend journal serialization format - The old version field is now reserved and always stores zero and ignores it when reading - This allows old entries to be used by new code without having to migrate the entire spend journal - Remove version field from gettxout RPC result - Convert UtxoEntry to represent a specific utxo instead of a transaction with all remaining utxos - Optimize for memory usage with an eye towards a utxo cache - Combine details such as whether the txout was contained in a coinbase, is spent, and is modified into a single packed field of bit flags - Align entry fields to eliminate extra padding since ultimately there will be a lot of these in memory - Introduce a free list for serializing an outpoint to the database key format to significantly reduce pressure on the GC - Update all related functions that previously dealt with transaction hashes to accept outpoints instead - Update all callers accordingly - Only add individually requested outputs from the mempool when constructing a mempool view - Modify the spend journal to always store the block height and coinbase information with every spent txout - Introduce code to handle fetching the missing information from another utxo from the same transaction in the event an old style entry is encountered - Make use of a database cursor with seek to do this much more efficiently than testing every possible output - Always decompress data loaded from the database now that a utxo entry only consists of a specific output - Introduce upgrade code to migrate the utxo set to the new format - Store versions of the utxoset and spend journal buckets - Allow migration process to be interrupted and resumed - Update all tests to expect the correct encodings, remove tests that no longer apply, and add new ones for the new expected behavior - Convert old tests for the legacy utxo format deserialization code to test the new function that is used during upgrade - Update the utxostore test data and add function that was used to convert it - Introduce a few new functions on UtxoViewpoint - AddTxOut for adding an individual txout versus all of them - addTxOut to handle the common code between the new AddTxOut and existing AddTxOuts - RemoveEntry for removing an individual txout - fetchEntryByHash for fetching any remaining utxo for a given transaction hash
2017-09-03 09:59:15 +02:00
hashCopy := outpoint.Hash
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
missingParents = append(missingParents, &hashCopy)
}
}
if len(missingParents) > 0 {
return missingParents, nil, nil
}
// Don't allow the transaction into the mempool unless its sequence
// lock is active, meaning that it'll be allowed into the next block
// with respect to its defined relative lock times.
sequenceLock, err := mp.cfg.CalcSequenceLock(tx, utxoView)
if err != nil {
if cerr, ok := err.(blockchain.RuleError); ok {
return nil, nil, chainRuleError(cerr)
}
return nil, nil, err
}
if !blockchain.SequenceLockActive(sequenceLock, nextBlockHeight,
medianTimePast) {
return nil, nil, txRuleError(wire.RejectNonstandard,
"transaction's sequence locks on inputs not met")
}
// Perform several checks on the transaction inputs using the invariant
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
// rules in blockchain for what transactions are allowed into blocks.
// Also returns the fees associated with the transaction which will be
// used later.
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
txFee, err := blockchain.CheckTransactionInputs(tx, nextBlockHeight,
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
utxoView, mp.cfg.ChainParams)
if err != nil {
if cerr, ok := err.(blockchain.RuleError); ok {
return nil, nil, chainRuleError(cerr)
}
return nil, nil, err
}
// Don't allow transactions with non-standard inputs if the network
// parameters forbid their acceptance.
if !mp.cfg.Policy.AcceptNonStd {
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
err := checkInputsStandard(tx, utxoView)
if err != nil {
// Attempt to extract a reject code from the error so
// it can be retained. When not possible, fall back to
// a non standard error.
rejectCode, found := extractRejectCode(err)
if !found {
rejectCode = wire.RejectNonstandard
}
str := fmt.Sprintf("transaction %v has a non-standard "+
"input: %v", txHash, err)
return nil, nil, txRuleError(rejectCode, str)
}
}
// NOTE: if you modify this code to accept non-standard transactions,
// you should add code here to check that the transaction does a
// reasonable number of ECDSA signature verifications.
// Don't allow transactions with an excessive number of signature
// operations which would result in making it impossible to mine. Since
// the coinbase address itself can contain signature operations, the
// maximum allowed signature operations per transaction is less than
// the maximum allowed signature operations per block.
// TODO(roasbeef): last bool should be conditional on segwit activation
sigOpCost, err := blockchain.GetSigOpCost(tx, false, utxoView, true, true)
if err != nil {
if cerr, ok := err.(blockchain.RuleError); ok {
return nil, nil, chainRuleError(cerr)
}
return nil, nil, err
}
if sigOpCost > mp.cfg.Policy.MaxSigOpCostPerTx {
str := fmt.Sprintf("transaction %v sigop cost is too high: %d > %d",
txHash, sigOpCost, mp.cfg.Policy.MaxSigOpCostPerTx)
return nil, nil, txRuleError(wire.RejectNonstandard, str)
}
// Don't allow transactions with fees too low to get into a mined block.
//
// Most miners allow a free transaction area in blocks they mine to go
// alongside the area used for high-priority transactions as well as
// transactions with fees. A transaction size of up to 1000 bytes is
// considered safe to go into this section. Further, the minimum fee
// calculated below on its own would encourage several small
// transactions to avoid fees rather than one single larger transaction
// which is more desirable. Therefore, as long as the size of the
// transaction does not exceeed 1000 less than the reserved space for
// high-priority transactions, don't require a fee for it.
serializedSize := GetTxVirtualSize(tx)
minFee := calcMinRequiredTxRelayFee(serializedSize,
mp.cfg.Policy.MinRelayTxFee)
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
if serializedSize >= (DefaultBlockPrioritySize-1000) && txFee < minFee {
str := fmt.Sprintf("transaction %v has %d fees which is under "+
"the required amount of %d", txHash, txFee,
minFee)
return nil, nil, txRuleError(wire.RejectInsufficientFee, str)
}
// Require that free transactions have sufficient priority to be mined
// in the next block. Transactions which are being added back to the
// memory pool from blocks that have been disconnected during a reorg
// are exempted.
if isNew && !mp.cfg.Policy.DisableRelayPriority && txFee < minFee {
currentPriority := mining.CalcPriority(tx.MsgTx(), utxoView,
nextBlockHeight)
if currentPriority <= mining.MinHighPriority {
str := fmt.Sprintf("transaction %v has insufficient "+
"priority (%g <= %g)", txHash,
currentPriority, mining.MinHighPriority)
return nil, nil, txRuleError(wire.RejectInsufficientFee, str)
}
}
// Free-to-relay transactions are rate limited here to prevent
// penny-flooding with tiny transactions as a form of attack.
if rateLimit && txFee < minFee {
nowUnix := time.Now().Unix()
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
// Decay passed data with an exponentially decaying ~10 minute
// window - matches bitcoind handling.
mp.pennyTotal *= math.Pow(1.0-1.0/600.0,
float64(nowUnix-mp.lastPennyUnix))
mp.lastPennyUnix = nowUnix
// Are we still over the limit?
if mp.pennyTotal >= mp.cfg.Policy.FreeTxRelayLimit*10*1000 {
str := fmt.Sprintf("transaction %v has been rejected "+
"by the rate limiter due to low fees", txHash)
return nil, nil, txRuleError(wire.RejectInsufficientFee, str)
}
oldTotal := mp.pennyTotal
mp.pennyTotal += float64(serializedSize)
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
log.Tracef("rate limit: curTotal %v, nextTotal: %v, "+
"limit %v", oldTotal, mp.pennyTotal,
mp.cfg.Policy.FreeTxRelayLimit*10*1000)
}
// If the transaction has any conflicts and we've made it this far, then
// we're processing a potential replacement.
var conflicts map[chainhash.Hash]*btcutil.Tx
if isReplacement {
conflicts, err = mp.validateReplacement(tx, txFee)
if err != nil {
return nil, nil, err
}
}
// Verify crypto signatures for each input and reject the transaction if
// any don't verify.
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
err = blockchain.ValidateTransactionScripts(tx, utxoView,
txscript.StandardVerifyFlags, mp.cfg.SigCache,
mp.cfg.HashCache)
if err != nil {
if cerr, ok := err.(blockchain.RuleError); ok {
return nil, nil, chainRuleError(cerr)
}
return nil, nil, err
}
// Now that we've deemed the transaction as valid, we can add it to the
// mempool. If it ended up replacing any transactions, we'll remove them
// first.
for _, conflict := range conflicts {
log.Debugf("Replacing transaction %v (fee_rate=%v sat/kb) "+
"with %v (fee_rate=%v sat/kb)\n", conflict.Hash(),
mp.pool[*conflict.Hash()].FeePerKB, tx.Hash(),
txFee*1000/serializedSize)
// The conflict set should already include the descendants for
// each one, so we don't need to remove the redeemers within
// this call as they'll be removed eventually.
mp.removeTransaction(conflict, false)
}
txD := mp.addTransaction(utxoView, tx, bestHeight, txFee)
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
log.Debugf("Accepted transaction %v (pool size: %v)", txHash,
len(mp.pool))
return nil, txD, nil
}
// MaybeAcceptTransaction is the main workhorse for handling insertion of new
// free-standing transactions into a memory pool. It includes functionality
// such as rejecting duplicate transactions, ensuring transactions follow all
// rules, detecting orphan transactions, and insertion into the memory pool.
//
// If the transaction is an orphan (missing parent transactions), the
// transaction is NOT added to the orphan pool, but each unknown referenced
// parent is returned. Use ProcessTransaction instead if new orphans should
// be added to the orphan pool.
//
// This function is safe for concurrent access.
func (mp *TxPool) MaybeAcceptTransaction(tx *btcutil.Tx, isNew, rateLimit bool) ([]*chainhash.Hash, *TxDesc, error) {
// Protect concurrent access.
mp.mtx.Lock()
hashes, txD, err := mp.maybeAcceptTransaction(tx, isNew, rateLimit, true)
mp.mtx.Unlock()
return hashes, txD, err
}
// processOrphans is the internal function which implements the public
// ProcessOrphans. See the comment for ProcessOrphans for more details.
//
// This function MUST be called with the mempool lock held (for writes).
func (mp *TxPool) processOrphans(acceptedTx *btcutil.Tx) []*TxDesc {
var acceptedTxns []*TxDesc
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
// Start with processing at least the passed transaction.
processList := list.New()
processList.PushBack(acceptedTx)
for processList.Len() > 0 {
// Pop the transaction to process from the front of the list.
firstElement := processList.Remove(processList.Front())
processItem := firstElement.(*btcutil.Tx)
prevOut := wire.OutPoint{Hash: *processItem.Hash()}
for txOutIdx := range processItem.MsgTx().TxOut {
// Look up all orphans that redeem the output that is
// now available. This will typically only be one, but
// it could be multiple if the orphan pool contains
// double spends. While it may seem odd that the orphan
// pool would allow this since there can only possibly
// ultimately be a single redeemer, it's important to
// track it this way to prevent malicious actors from
// being able to purposely constructing orphans that
// would otherwise make outputs unspendable.
//
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
// Skip to the next available output if there are none.
prevOut.Index = uint32(txOutIdx)
orphans, exists := mp.orphansByPrev[prevOut]
if !exists {
continue
}
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
// Potentially accept an orphan into the tx pool.
for _, tx := range orphans {
missing, txD, err := mp.maybeAcceptTransaction(
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
tx, true, true, false)
if err != nil {
// The orphan is now invalid, so there
// is no way any other orphans which
// redeem any of its outputs can be
// accepted. Remove them.
mp.removeOrphan(tx, true)
break
}
// Transaction is still an orphan. Try the next
// orphan which redeems this output.
if len(missing) > 0 {
continue
}
// Transaction was accepted into the main pool.
//
// Add it to the list of accepted transactions
// that are no longer orphans, remove it from
// the orphan pool, and add it to the list of
// transactions to process so any orphans that
// depend on it are handled too.
acceptedTxns = append(acceptedTxns, txD)
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
mp.removeOrphan(tx, false)
processList.PushBack(tx)
// Only one transaction for this outpoint can be
// accepted, so the rest are now double spends
// and are removed later.
break
}
}
}
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
// Recursively remove any orphans that also redeem any outputs redeemed
// by the accepted transactions since those are now definitive double
// spends.
mp.removeOrphanDoubleSpends(acceptedTx)
for _, txD := range acceptedTxns {
mp.removeOrphanDoubleSpends(txD.Tx)
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
}
return acceptedTxns
}
// ProcessOrphans determines if there are any orphans which depend on the passed
// transaction hash (it is possible that they are no longer orphans) and
// potentially accepts them to the memory pool. It repeats the process for the
// newly accepted transactions (to detect further orphans which may no longer be
// orphans) until there are no more.
//
// It returns a slice of transactions added to the mempool. A nil slice means
// no transactions were moved from the orphan pool to the mempool.
//
// This function is safe for concurrent access.
func (mp *TxPool) ProcessOrphans(acceptedTx *btcutil.Tx) []*TxDesc {
mp.mtx.Lock()
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
acceptedTxns := mp.processOrphans(acceptedTx)
mp.mtx.Unlock()
return acceptedTxns
}
// ProcessTransaction is the main workhorse for handling insertion of new
// free-standing transactions into the memory pool. It includes functionality
// such as rejecting duplicate transactions, ensuring transactions follow all
// rules, orphan transaction handling, and insertion into the memory pool.
//
// It returns a slice of transactions added to the mempool. When the
// error is nil, the list will include the passed transaction itself along
// with any additional orphan transaactions that were added as a result of
// the passed one being accepted.
//
// This function is safe for concurrent access.
func (mp *TxPool) ProcessTransaction(tx *btcutil.Tx, allowOrphan, rateLimit bool, tag Tag) ([]*TxDesc, error) {
log.Tracef("Processing transaction %v", tx.Hash())
// Protect concurrent access.
mp.mtx.Lock()
defer mp.mtx.Unlock()
// Potentially accept the transaction to the memory pool.
missingParents, txD, err := mp.maybeAcceptTransaction(tx, true, rateLimit,
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
true)
if err != nil {
return nil, err
}
if len(missingParents) == 0 {
// Accept any orphan transactions that depend on this
// transaction (they may no longer be orphans if all inputs
// are now available) and repeat for those accepted
// transactions until there are no more.
mempool: Stricter orphan evaluation and eviction. This modifies the way orphan removal and processing is done to more aggressively remove orphans that can no longer be valid due to other transactions being added or removed from the primary transaction pool. The net effect of these changes is that orphan pool will typically be much smaller which greatly improves its effectiveness. Previously, it would typically quickly reach the max allowed worst-case usage and effectively stay there forever. The following is a summary of the changes: - Modify the map that tracks which orphans redeem a given transaction to instead track by the specific outpoints that are redeemed - Modify the various orphan removal and processing functions to accept the full transaction rather than just its hash - Introduce a new flag on removeOrphans which specifies whether or not to remove the transactions that redeem the orphan being removed as well which is necessary since only some paths require it - Add a new function named removeOrphanDoubleSpends that is invoked whenever a transaction is added to the main pool and thus the outputs they spent become concrete spends - Introduce a new flag on maybeAcceptTransaction which specifies whether or not duplicate orphans should be rejected since only some paths require it - Modify processOrphans as follows: - Make use of the modified map - Use newly available flags and logic work more strictly work with tx chains - Recursively remove any orphans that also redeem any outputs redeemed by the accepted transactions - Several new tests to ensure proper functionality - Removing an orphan that doesn't exist is removed both when there is another orphan that redeems it and when there is not - Removing orphans works properly with orphan chains per the new remove redeemers flag - Removal of multi-input orphans that double spend an output when a concrete redeemer enters the transaction pool
2016-08-23 19:26:26 +02:00
newTxs := mp.processOrphans(tx)
acceptedTxs := make([]*TxDesc, len(newTxs)+1)
// Add the parent transaction first so remote nodes
// do not add orphans.
acceptedTxs[0] = txD
copy(acceptedTxs[1:], newTxs)
return acceptedTxs, nil
}
// The transaction is an orphan (has inputs missing). Reject
// it if the flag to allow orphans is not set.
if !allowOrphan {
// Only use the first missing parent transaction in
// the error message.
//
// NOTE: RejectDuplicate is really not an accurate
// reject code here, but it matches the reference
// implementation and there isn't a better choice due
// to the limited number of reject codes. Missing
// inputs is assumed to mean they are already spent
// which is not really always the case.
str := fmt.Sprintf("orphan transaction %v references "+
"outputs of unknown or fully-spent "+
"transaction %v", tx.Hash(), missingParents[0])
return nil, txRuleError(wire.RejectDuplicate, str)
}
// Potentially add the orphan transaction to the orphan pool.
err = mp.maybeAddOrphan(tx, tag)
return nil, err
}
// Count returns the number of transactions in the main pool. It does not
// include the orphan pool.
//
// This function is safe for concurrent access.
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
func (mp *TxPool) Count() int {
mp.mtx.RLock()
count := len(mp.pool)
mp.mtx.RUnlock()
return count
}
// TxHashes returns a slice of hashes for all of the transactions in the memory
2013-10-08 20:34:04 +02:00
// pool.
//
// This function is safe for concurrent access.
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
func (mp *TxPool) TxHashes() []*chainhash.Hash {
mp.mtx.RLock()
hashes := make([]*chainhash.Hash, len(mp.pool))
2013-10-08 07:04:51 +02:00
i := 0
for hash := range mp.pool {
hashCopy := hash
hashes[i] = &hashCopy
i++
}
mp.mtx.RUnlock()
2013-10-08 07:04:51 +02:00
return hashes
}
// TxDescs returns a slice of descriptors for all the transactions in the pool.
// The descriptors are to be treated as read only.
//
// This function is safe for concurrent access.
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
func (mp *TxPool) TxDescs() []*TxDesc {
mp.mtx.RLock()
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
descs := make([]*TxDesc, len(mp.pool))
i := 0
for _, desc := range mp.pool {
descs[i] = desc
i++
}
mp.mtx.RUnlock()
return descs
}
// MiningDescs returns a slice of mining descriptors for all the transactions
// in the pool.
//
// This is part of the mining.TxSource interface implementation and is safe for
// concurrent access as required by the interface contract.
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
func (mp *TxPool) MiningDescs() []*mining.TxDesc {
mp.mtx.RLock()
descs := make([]*mining.TxDesc, len(mp.pool))
i := 0
for _, desc := range mp.pool {
descs[i] = &desc.TxDesc
i++
}
mp.mtx.RUnlock()
return descs
}
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
// RawMempoolVerbose returns all of the entries in the mempool as a fully
// populated btcjson result.
//
// This function is safe for concurrent access.
func (mp *TxPool) RawMempoolVerbose() map[string]*btcjson.GetRawMempoolVerboseResult {
mp.mtx.RLock()
defer mp.mtx.RUnlock()
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
result := make(map[string]*btcjson.GetRawMempoolVerboseResult,
len(mp.pool))
bestHeight := mp.cfg.BestHeight()
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
for _, desc := range mp.pool {
// Calculate the current priority based on the inputs to
// the transaction. Use zero if one or more of the
// input transactions can't be found for some reason.
tx := desc.Tx
var currentPriority float64
utxos, err := mp.fetchInputUtxos(tx)
if err == nil {
currentPriority = mining.CalcPriority(tx.MsgTx(), utxos,
bestHeight+1)
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
}
mpd := &btcjson.GetRawMempoolVerboseResult{
Size: int32(tx.MsgTx().SerializeSize()),
Vsize: int32(GetTxVirtualSize(tx)),
Weight: int32(blockchain.GetTransactionWeight(tx)),
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
Fee: btcutil.Amount(desc.Fee).ToBTC(),
Time: desc.Added.Unix(),
Height: int64(desc.Height),
StartingPriority: desc.StartingPriority,
CurrentPriority: currentPriority,
Depends: make([]string, 0),
}
for _, txIn := range tx.MsgTx().TxIn {
hash := &txIn.PreviousOutPoint.Hash
if mp.haveTransaction(hash) {
mpd.Depends = append(mpd.Depends,
hash.String())
}
}
result[tx.Hash().String()] = mpd
}
return result
}
// LastUpdated returns the last time a transaction was added to or removed from
// the main pool. It does not include the orphan pool.
//
// This function is safe for concurrent access.
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
func (mp *TxPool) LastUpdated() time.Time {
2016-02-12 19:44:56 +01:00
return time.Unix(atomic.LoadInt64(&mp.lastUpdated), 0)
}
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
// New returns a new memory pool for validating and storing standalone
// transactions until they are mined into a block.
mempool: Refactor mempool code to its own package. (#737) This does the minimum work necessary to refactor the mempool code into its own package. The idea is that separating this code into its own package will greatly improve its testability, allow independent benchmarking and profiling, and open up some interesting opportunities for future development related to the memory pool. There are likely some areas related to policy that could be further refactored, however it is better to do that in future commits in order to keep the changeset as small as possible during this refactor. Overview of the major changes: - Create the new package - Move several files into the new package: - mempool.go -> mempool/mempool.go - mempoolerror.go -> mempool/error.go - policy.go -> mempool/policy.go - policy_test.go -> mempool/policy_test.go - Update mempool logging to use the new mempool package logger - Rename mempoolPolicy to Policy (so it's now mempool.Policy) - Rename mempoolConfig to Config (so it's now mempool.Config) - Rename mempoolTxDesc to TxDesc (so it's now mempool.TxDesc) - Rename txMemPool to TxPool (so it's now mempool.TxPool) - Move defaultBlockPrioritySize to the new package and export it - Export DefaultMinRelayTxFee from the mempool package - Export the CalcPriority function from the mempool package - Introduce a new RawMempoolVerbose function on the TxPool and update the RPC server to use it - Update all references to the mempool to use the package. - Add a skeleton README.md
2016-08-19 18:08:37 +02:00
func New(cfg *Config) *TxPool {
return &TxPool{
cfg: *cfg,
pool: make(map[chainhash.Hash]*TxDesc),
orphans: make(map[chainhash.Hash]*orphanTx),
orphansByPrev: make(map[wire.OutPoint]map[chainhash.Hash]*btcutil.Tx),
nextExpireScan: time.Now().Add(orphanExpireScanInterval),
outpoints: make(map[wire.OutPoint]*btcutil.Tx),
}
}