lbcd/server.go

3283 lines
101 KiB
Go
Raw Normal View History

// Copyright (c) 2013-2017 The btcsuite developers
// Copyright (c) 2015-2018 The Decred developers
2013-08-06 23:55:22 +02:00
// Use of this source code is governed by an ISC
// license that can be found in the LICENSE file.
package main
import (
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
"bytes"
"crypto/rand"
"crypto/tls"
"encoding/binary"
"errors"
"fmt"
"math"
2013-08-06 23:55:22 +02:00
"net"
"runtime"
"sort"
"strconv"
"strings"
2013-08-06 23:55:22 +02:00
"sync"
"sync/atomic"
2013-08-06 23:55:22 +02:00
"time"
2014-07-02 15:50:08 +02:00
"github.com/btcsuite/btcd/addrmgr"
"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"
"github.com/btcsuite/btcd/chaincfg"
"github.com/btcsuite/btcd/chaincfg/chainhash"
"github.com/btcsuite/btcd/connmgr"
"github.com/btcsuite/btcd/database"
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/mempool"
"github.com/btcsuite/btcd/mining"
"github.com/btcsuite/btcd/mining/cpuminer"
"github.com/btcsuite/btcd/netsync"
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
"github.com/btcsuite/btcd/peer"
Integrate a valid ECDSA signature cache into btcd Introduce an ECDSA signature verification into btcd in order to mitigate a certain DoS attack and as a performance optimization. The benefits of SigCache are two fold. Firstly, usage of SigCache mitigates a DoS attack wherein an attacker causes a victim's client to hang due to worst-case behavior triggered while processing attacker crafted invalid transactions. A detailed description of the mitigated DoS attack can be found here: https://bitslog.wordpress.com/2013/01/23/fixed-bitcoin-vulnerability-explanation-why-the-signature-cache-is-a-dos-protection/ Secondly, usage of the SigCache introduces a signature verification optimization which speeds up the validation of transactions within a block, if they've already been seen and verified within the mempool. The server itself manages the sigCache instance. The blockManager and txMempool respectively now receive pointers to the created sigCache instance. All read (sig triplet existence) operations on the sigCache will not block unless a separate goroutine is adding an entry (writing) to the sigCache. GetBlockTemplate generation now also utilizes the sigCache in order to avoid unnecessarily double checking signatures when generating a template after previously accepting a txn to the mempool. Consequently, the CPU miner now also employs the same optimization. The maximum number of entries for the sigCache has been introduced as a config parameter in order to allow users to configure the amount of memory consumed by this new additional caching.
2015-09-25 01:22:00 +02:00
"github.com/btcsuite/btcd/txscript"
"github.com/btcsuite/btcd/wire"
"github.com/btcsuite/btcutil"
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
"github.com/btcsuite/btcutil/bloom"
2013-08-06 23:55:22 +02:00
)
const (
// defaultServices describes the default services that are supported by
// the server.
defaultServices = wire.SFNodeNetwork | wire.SFNodeBloom |
wire.SFNodeWitness | wire.SFNodeCF
2013-08-06 23:55:22 +02:00
// defaultRequiredServices describes the default services that are
// required to be supported by outbound peers.
defaultRequiredServices = wire.SFNodeNetwork
// defaultTargetOutbound is the default number of outbound peers to target.
defaultTargetOutbound = 8
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// connectionRetryInterval is the base amount of time to wait in between
// retries when connecting to persistent peers. It is adjusted by the
// number of retries such that there is a retry backoff.
connectionRetryInterval = time.Second * 5
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
)
var (
// userAgentName is the user agent name and is used to help identify
// ourselves to other bitcoin peers.
userAgentName = "LBRY.GO"
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// userAgentVersion is the user agent version and is used to help
// identify ourselves to other bitcoin peers.
userAgentVersion = fmt.Sprintf("%d.%d.%d", appMajor, appMinor, appPatch)
)
// zeroHash is the zero value hash (all zeros). It is defined as a convenience.
var zeroHash chainhash.Hash
// onionAddr implements the net.Addr interface and represents a tor address.
type onionAddr struct {
addr string
}
// String returns the onion address.
//
// This is part of the net.Addr interface.
func (oa *onionAddr) String() string {
return oa.addr
}
// Network returns "onion".
//
// This is part of the net.Addr interface.
func (oa *onionAddr) Network() string {
return "onion"
}
// Ensure onionAddr implements the net.Addr interface.
var _ net.Addr = (*onionAddr)(nil)
2018-09-08 16:34:41 +02:00
// simpleAddr implements the net.Addr interface with two struct fields
type simpleAddr struct {
net, addr string
}
// String returns the address.
//
// This is part of the net.Addr interface.
func (a simpleAddr) String() string {
return a.addr
}
// Network returns the network.
//
// This is part of the net.Addr interface.
func (a simpleAddr) Network() string {
return a.net
}
// Ensure simpleAddr implements the net.Addr interface.
var _ net.Addr = simpleAddr{}
2013-08-06 23:55:22 +02:00
// broadcastMsg provides the ability to house a bitcoin message to be broadcast
// to all connected peers except specified excluded peers.
type broadcastMsg struct {
message wire.Message
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
excludePeers []*serverPeer
2013-08-06 23:55:22 +02:00
}
// broadcastInventoryAdd is a type used to declare that the InvVect it contains
// needs to be added to the rebroadcast map
type broadcastInventoryAdd relayMsg
// broadcastInventoryDel is a type used to declare that the InvVect it contains
// needs to be removed from the rebroadcast map
type broadcastInventoryDel *wire.InvVect
// relayMsg packages an inventory vector along with the newly discovered
// inventory so the relay has access to that information.
type relayMsg struct {
invVect *wire.InvVect
data interface{}
}
// updatePeerHeightsMsg is a message sent from the blockmanager to the server
// after a new block has been accepted. The purpose of the message is to update
// the heights of peers that were known to announce the block before we
// connected it to the main chain or recognized it as an orphan. With these
// updates, peer heights will be kept up to date, allowing for fresh data when
// selecting sync peer candidacy.
type updatePeerHeightsMsg struct {
newHash *chainhash.Hash
newHeight int32
originPeer *peer.Peer
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
}
// peerState maintains state of inbound, persistent, outbound peers as well
// as banned peers and outbound groups.
type peerState struct {
inboundPeers map[int32]*serverPeer
outboundPeers map[int32]*serverPeer
persistentPeers map[int32]*serverPeer
banned map[string]time.Time
outboundGroups map[string]int
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
}
// Count returns the count of all known peers.
func (ps *peerState) Count() int {
return len(ps.inboundPeers) + len(ps.outboundPeers) +
len(ps.persistentPeers)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
}
// forAllOutboundPeers is a helper function that runs closure on all outbound
// peers known to peerState.
func (ps *peerState) forAllOutboundPeers(closure func(sp *serverPeer)) {
for _, e := range ps.outboundPeers {
closure(e)
}
for _, e := range ps.persistentPeers {
closure(e)
}
}
// forAllPeers is a helper function that runs closure on all peers known to
// peerState.
func (ps *peerState) forAllPeers(closure func(sp *serverPeer)) {
for _, e := range ps.inboundPeers {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
closure(e)
}
ps.forAllOutboundPeers(closure)
}
// cfHeaderKV is a tuple of a filter header and its associated block hash. The
// struct is used to cache cfcheckpt responses.
type cfHeaderKV struct {
blockHash chainhash.Hash
filterHeader chainhash.Hash
}
2013-08-06 23:55:22 +02:00
// server provides a bitcoin server for handling communications to and from
// bitcoin peers.
type server struct {
// The following variables must only be used atomically.
// Putting the uint64s first makes them 64-bit aligned for 32-bit systems.
bytesReceived uint64 // Total bytes received from all peers since start.
bytesSent uint64 // Total bytes sent by all peers since start.
started int32
shutdown int32
shutdownSched int32
chainParams *chaincfg.Params
addrManager *addrmgr.AddrManager
connManager *connmgr.ConnManager
Integrate a valid ECDSA signature cache into btcd Introduce an ECDSA signature verification into btcd in order to mitigate a certain DoS attack and as a performance optimization. The benefits of SigCache are two fold. Firstly, usage of SigCache mitigates a DoS attack wherein an attacker causes a victim's client to hang due to worst-case behavior triggered while processing attacker crafted invalid transactions. A detailed description of the mitigated DoS attack can be found here: https://bitslog.wordpress.com/2013/01/23/fixed-bitcoin-vulnerability-explanation-why-the-signature-cache-is-a-dos-protection/ Secondly, usage of the SigCache introduces a signature verification optimization which speeds up the validation of transactions within a block, if they've already been seen and verified within the mempool. The server itself manages the sigCache instance. The blockManager and txMempool respectively now receive pointers to the created sigCache instance. All read (sig triplet existence) operations on the sigCache will not block unless a separate goroutine is adding an entry (writing) to the sigCache. GetBlockTemplate generation now also utilizes the sigCache in order to avoid unnecessarily double checking signatures when generating a template after previously accepting a txn to the mempool. Consequently, the CPU miner now also employs the same optimization. The maximum number of entries for the sigCache has been introduced as a config parameter in order to allow users to configure the amount of memory consumed by this new additional caching.
2015-09-25 01:22:00 +02:00
sigCache *txscript.SigCache
hashCache *txscript.HashCache
rpcServer *rpcServer
syncManager *netsync.SyncManager
chain *blockchain.BlockChain
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
txMemPool *mempool.TxPool
cpuMiner *cpuminer.CPUMiner
modifyRebroadcastInv chan interface{}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
newPeers chan *serverPeer
donePeers chan *serverPeer
banPeers chan *serverPeer
query chan interface{}
relayInv chan relayMsg
broadcast chan broadcastMsg
peerHeightsUpdate chan updatePeerHeightsMsg
wg sync.WaitGroup
quit chan struct{}
nat NAT
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
db database.DB
timeSource blockchain.MedianTimeSource
services wire.ServiceFlag
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
// The following fields are used for optional indexes. They will be nil
// if the associated index is not enabled. These fields are set during
// initial creation of the server and never changed afterwards, so they
// do not need to be protected for concurrent access.
txIndex *indexers.TxIndex
addrIndex *indexers.AddrIndex
cfIndex *indexers.CfIndex
// The fee estimator keeps track of how long transactions are left in
// the mempool before they are mined into blocks.
feeEstimator *mempool.FeeEstimator
// cfCheckptCaches stores a cached slice of filter headers for cfcheckpt
// messages for each filter type.
cfCheckptCaches map[wire.FilterType][]cfHeaderKV
cfCheckptCachesMtx sync.RWMutex
// agentBlacklist is a list of blacklisted substrings by which to filter
// user agents.
agentBlacklist []string
// agentWhitelist is a list of whitelisted user agent substrings, no
// whitelisting will be applied if the list is empty or nil.
agentWhitelist []string
2013-08-06 23:55:22 +02:00
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// serverPeer extends the peer to maintain state shared by the server and
// the blockmanager.
type serverPeer struct {
// The following variables must only be used atomically
feeFilter int64
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
*peer.Peer
connReq *connmgr.ConnReq
server *server
persistent bool
continueHash *chainhash.Hash
relayMtx sync.Mutex
disableRelayTx bool
sentAddrs bool
2017-08-31 15:59:43 +02:00
isWhitelisted bool
filter *bloom.Filter
addressesMtx sync.RWMutex
knownAddresses map[string]struct{}
banScore connmgr.DynamicBanScore
quit chan struct{}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// The following chans are used to sync blockmanager and server.
txProcessed chan struct{}
blockProcessed chan struct{}
}
// newServerPeer returns a new serverPeer instance. The peer needs to be set by
// the caller.
func newServerPeer(s *server, isPersistent bool) *serverPeer {
return &serverPeer{
server: s,
persistent: isPersistent,
filter: bloom.LoadFilter(nil),
knownAddresses: make(map[string]struct{}),
quit: make(chan struct{}),
txProcessed: make(chan struct{}, 1),
blockProcessed: make(chan struct{}, 1),
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +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
// newestBlock returns the current best block hash and height using the format
// required by the configuration for the peer package.
func (sp *serverPeer) newestBlock() (*chainhash.Hash, int32, error) {
best := sp.server.chain.BestSnapshot()
return &best.Hash, best.Height, nil
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
}
// addKnownAddresses adds the given addresses to the set of known addresses to
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// the peer to prevent sending duplicate addresses.
func (sp *serverPeer) addKnownAddresses(addresses []*wire.NetAddress) {
sp.addressesMtx.Lock()
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
for _, na := range addresses {
sp.knownAddresses[addrmgr.NetAddressKey(na)] = struct{}{}
}
sp.addressesMtx.Unlock()
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
}
// addressKnown true if the given address is already known to the peer.
func (sp *serverPeer) addressKnown(na *wire.NetAddress) bool {
sp.addressesMtx.RLock()
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
_, exists := sp.knownAddresses[addrmgr.NetAddressKey(na)]
sp.addressesMtx.RUnlock()
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
return exists
}
// setDisableRelayTx toggles relaying of transactions for the given peer.
// It is safe for concurrent access.
func (sp *serverPeer) setDisableRelayTx(disable bool) {
sp.relayMtx.Lock()
sp.disableRelayTx = disable
sp.relayMtx.Unlock()
}
// relayTxDisabled returns whether or not relaying of transactions for the given
// peer is disabled.
// It is safe for concurrent access.
func (sp *serverPeer) relayTxDisabled() bool {
sp.relayMtx.Lock()
isDisabled := sp.disableRelayTx
sp.relayMtx.Unlock()
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
return isDisabled
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
}
// pushAddrMsg sends an addr message to the connected peer using the provided
// addresses.
func (sp *serverPeer) pushAddrMsg(addresses []*wire.NetAddress) {
// Filter addresses already known to the peer.
addrs := make([]*wire.NetAddress, 0, len(addresses))
for _, addr := range addresses {
if !sp.addressKnown(addr) {
addrs = append(addrs, addr)
}
}
known, err := sp.PushAddrMsg(addrs)
if err != nil {
peerLog.Errorf("Can't push address message to %s: %v", sp.Peer, err)
sp.Disconnect()
return
}
sp.addKnownAddresses(known)
}
// addBanScore increases the persistent and decaying ban score fields by the
// values passed as parameters. If the resulting score exceeds half of the ban
// threshold, a warning is logged including the reason provided. Further, if
// the score is above the ban threshold, the peer will be banned and
// disconnected.
func (sp *serverPeer) addBanScore(persistent, transient uint32, reason string) bool {
// No warning is logged and no score is calculated if banning is disabled.
if cfg.DisableBanning {
return false
}
2017-08-31 15:59:43 +02:00
if sp.isWhitelisted {
peerLog.Debugf("Misbehaving whitelisted peer %s: %s", sp, reason)
return false
2017-08-31 15:59:43 +02:00
}
warnThreshold := cfg.BanThreshold >> 1
if transient == 0 && persistent == 0 {
// The score is not being increased, but a warning message is still
// logged if the score is above the warn threshold.
score := sp.banScore.Int()
if score > warnThreshold {
peerLog.Warnf("Misbehaving peer %s: %s -- ban score is %d, "+
"it was not increased this time", sp, reason, score)
}
return false
}
score := sp.banScore.Increase(persistent, transient)
if score > warnThreshold {
peerLog.Warnf("Misbehaving peer %s: %s -- ban score increased to %d", sp, reason, score)
if score > cfg.BanThreshold {
if sp.server.ConnectedCount() <= 1 {
peerLog.Warnf("Refusing to ban peer %s as it is the only peer", sp)
return false
}
peerLog.Warnf("Misbehaving peer %s -- banning and disconnecting", sp)
sp.server.BanPeer(sp)
sp.Disconnect()
return true
}
}
return false
}
// hasServices returns whether or not the provided advertised service flags have
// all of the provided desired service flags set.
func hasServices(advertised, desired wire.ServiceFlag) bool {
return advertised&desired == desired
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// OnVersion is invoked when a peer receives a version bitcoin message
// and is used to negotiate the protocol version details as well as kick start
// the communications.
func (sp *serverPeer) OnVersion(_ *peer.Peer, msg *wire.MsgVersion) *wire.MsgReject {
// Update the address manager with the advertised services for outbound
// connections in case they have changed. This is not done for inbound
// connections to help prevent malicious behavior and is skipped when
// running on the simulation test network since it is only intended to
// connect to specified peers and actively avoids advertising and
// connecting to discovered peers.
//
// NOTE: This is done before rejecting peers that are too old to ensure
// it is updated regardless in the case a new minimum protocol version is
// enforced and the remote node has not upgraded yet.
isInbound := sp.Inbound()
remoteAddr := sp.NA()
addrManager := sp.server.addrManager
if !cfg.SimNet && !isInbound {
addrManager.SetServices(remoteAddr, msg.Services)
}
// Ignore peers that have a protcol version that is too old. The peer
// negotiation logic will disconnect it after this callback returns.
if msg.ProtocolVersion < int32(peer.MinAcceptableProtocolVersion) {
return nil
}
// Reject outbound peers that are not full nodes.
wantServices := wire.SFNodeNetwork
if !isInbound && !hasServices(msg.Services, wantServices) {
missingServices := wantServices & ^msg.Services
srvrLog.Debugf("Rejecting peer %s with services %v due to not "+
"providing desired services %v", sp.Peer, msg.Services,
missingServices)
reason := fmt.Sprintf("required services %#x not offered",
uint64(missingServices))
return wire.NewMsgReject(msg.Command(), wire.RejectNonstandard, reason)
}
if !cfg.SimNet && !isInbound {
// After soft-fork activation, only make outbound
// connection to peers if they flag that they're segwit
// enabled.
chain := sp.server.chain
segwitActive, err := chain.IsDeploymentActive(chaincfg.DeploymentSegwit)
if err != nil {
peerLog.Errorf("Unable to query for segwit soft-fork state: %v",
err)
return nil
}
if segwitActive && !sp.IsWitnessEnabled() {
peerLog.Infof("Disconnecting non-segwit peer %v, isn't segwit "+
"enabled and we need more segwit enabled peers", sp)
sp.Disconnect()
return nil
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
}
// Add the remote peer time as a sample for creating an offset against
// the local clock to keep the network time in sync.
sp.server.timeSource.AddTimeSample(sp.Addr(), msg.Timestamp)
// Choose whether or not to relay transactions before a filter command
// is received.
sp.setDisableRelayTx(msg.DisableRelayTx)
return nil
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
}
// OnVerAck is invoked when a peer receives a verack bitcoin message and is used
// to kick start communication with them.
func (sp *serverPeer) OnVerAck(_ *peer.Peer, _ *wire.MsgVerAck) {
sp.server.AddPeer(sp)
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// OnMemPool is invoked when a peer receives a mempool bitcoin message.
// It creates and sends an inventory message with the contents of the memory
// pool up to the maximum inventory allowed per message. When the peer has a
// bloom filter loaded, the contents are filtered accordingly.
func (sp *serverPeer) OnMemPool(_ *peer.Peer, msg *wire.MsgMemPool) {
// Only allow mempool requests if the server has bloom filtering
// enabled.
if sp.server.services&wire.SFNodeBloom != wire.SFNodeBloom {
peerLog.Debugf("peer %v sent mempool request with bloom "+
"filtering disabled -- disconnecting", sp)
sp.Disconnect()
return
}
// A decaying ban score increase is applied to prevent flooding.
// The ban score accumulates and passes the ban threshold if a burst of
// mempool messages comes from a peer. The score decays each minute to
// half of its value.
if sp.addBanScore(0, 33, "mempool") {
return
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// Generate inventory message with the available transactions in the
// transaction memory pool. Limit it to the max allowed inventory
// per message. The NewMsgInvSizeHint function automatically limits
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// the passed hint to the maximum allowed, so it's safe to pass it
// without double checking it here.
txMemPool := sp.server.txMemPool
txDescs := txMemPool.TxDescs()
invMsg := wire.NewMsgInvSizeHint(uint(len(txDescs)))
for _, txDesc := range txDescs {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// Either add all transactions when there is no bloom filter,
// or only the transactions that match the filter when there is
// one.
if !sp.filter.IsLoaded() || sp.filter.MatchTxAndUpdate(txDesc.Tx) {
iv := wire.NewInvVect(wire.InvTypeTx, txDesc.Tx.Hash())
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
invMsg.AddInvVect(iv)
if len(invMsg.InvList)+1 > wire.MaxInvPerMsg {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
break
}
}
}
// Send the inventory message if there is anything to send.
if len(invMsg.InvList) > 0 {
sp.QueueMessage(invMsg, nil)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
}
}
// OnTx is invoked when a peer receives a tx bitcoin message. It blocks
// until the bitcoin transaction has been fully processed. Unlock the block
// handler this does not serialize all transactions through a single thread
// transactions don't rely on the previous one in a linear fashion like blocks.
func (sp *serverPeer) OnTx(_ *peer.Peer, msg *wire.MsgTx) {
if cfg.BlocksOnly {
peerLog.Tracef("Ignoring tx %v from %v - blocksonly enabled",
msg.TxHash(), sp)
return
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// Add the transaction to the known inventory for the peer.
// Convert the raw MsgTx to a btcutil.Tx which provides some convenience
// methods and things such as hash caching.
tx := btcutil.NewTx(msg)
iv := wire.NewInvVect(wire.InvTypeTx, tx.Hash())
sp.AddKnownInventory(iv)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// Queue the transaction up to be handled by the sync manager and
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// intentionally block further receives until the transaction is fully
// processed and known good or bad. This helps prevent a malicious peer
// from queuing up a bunch of bad transactions before disconnecting (or
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// being disconnected) and wasting memory.
sp.server.syncManager.QueueTx(tx, sp.Peer, sp.txProcessed)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
<-sp.txProcessed
}
// OnBlock is invoked when a peer receives a block bitcoin message. It
// blocks until the bitcoin block has been fully processed.
func (sp *serverPeer) OnBlock(_ *peer.Peer, msg *wire.MsgBlock, buf []byte) {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// Convert the raw MsgBlock to a btcutil.Block which provides some
// convenience methods and things such as hash caching.
block := btcutil.NewBlockFromBlockAndBytes(msg, buf)
// Add the block to the known inventory for the peer.
iv := wire.NewInvVect(wire.InvTypeBlock, block.Hash())
sp.AddKnownInventory(iv)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// Queue the block up to be handled by the block
// manager and intentionally block further receives
// until the bitcoin block is fully processed and known
// good or bad. This helps prevent a malicious peer
// from queuing up a bunch of bad blocks before
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// disconnecting (or being disconnected) and wasting
// memory. Additionally, this behavior is depended on
// by at least the block acceptance test tool as the
// reference implementation processes blocks in the same
// thread and therefore blocks further messages until
// the bitcoin block has been fully processed.
sp.server.syncManager.QueueBlock(block, sp.Peer, sp.blockProcessed)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
<-sp.blockProcessed
}
// OnInv is invoked when a peer receives an inv bitcoin message and is
// used to examine the inventory being advertised by the remote peer and react
// accordingly. We pass the message down to blockmanager which will call
// QueueMessage with any appropriate responses.
func (sp *serverPeer) OnInv(_ *peer.Peer, msg *wire.MsgInv) {
if !cfg.BlocksOnly {
if len(msg.InvList) > 0 {
sp.server.syncManager.QueueInv(msg, sp.Peer)
}
return
}
newInv := wire.NewMsgInvSizeHint(uint(len(msg.InvList)))
for _, invVect := range msg.InvList {
if invVect.Type == wire.InvTypeTx {
peerLog.Tracef("Ignoring tx %v in inv from %v -- "+
"blocksonly enabled", invVect.Hash, sp)
if sp.ProtocolVersion() >= wire.BIP0037Version {
peerLog.Infof("Peer %v is announcing "+
"transactions -- disconnecting", sp)
sp.Disconnect()
return
}
continue
}
err := newInv.AddInvVect(invVect)
if err != nil {
peerLog.Errorf("Failed to add inventory vector: %v", err)
break
}
}
if len(newInv.InvList) > 0 {
sp.server.syncManager.QueueInv(newInv, sp.Peer)
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
}
// OnHeaders is invoked when a peer receives a headers bitcoin
// message. The message is passed down to the sync manager.
func (sp *serverPeer) OnHeaders(_ *peer.Peer, msg *wire.MsgHeaders) {
sp.server.syncManager.QueueHeaders(msg, sp.Peer)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
}
// handleGetData is invoked when a peer receives a getdata bitcoin message and
// is used to deliver block and transaction information.
func (sp *serverPeer) OnGetData(_ *peer.Peer, msg *wire.MsgGetData) {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
numAdded := 0
notFound := wire.NewMsgNotFound()
length := len(msg.InvList)
// A decaying ban score increase is applied to prevent exhausting resources
// with unusually large inventory queries.
// Requesting more than the maximum inventory vector length within a short
// period of time yields a score above the default ban threshold. Sustained
// bursts of small requests are not penalized as that would potentially ban
// peers performing IBD.
// This incremental score decays each minute to half of its value.
if sp.addBanScore(0, uint32(length)*99/wire.MaxInvPerMsg, "getdata") {
return
}
// We wait on this wait channel periodically to prevent queuing
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// far more data than we can send in a reasonable time, wasting memory.
// The waiting occurs after the database fetch for the next one to
// provide a little pipelining.
var waitChan chan struct{}
doneChan := make(chan struct{}, 1)
for i, iv := range msg.InvList {
var c chan struct{}
// If this will be the last message we send.
if i == length-1 && len(notFound.InvList) == 0 {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
c = doneChan
} else if (i+1)%3 == 0 {
// Buffered so as to not make the send goroutine block.
c = make(chan struct{}, 1)
}
var err error
switch iv.Type {
case wire.InvTypeWitnessTx:
err = sp.server.pushTxMsg(sp, &iv.Hash, c, waitChan, wire.WitnessEncoding)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
case wire.InvTypeTx:
err = sp.server.pushTxMsg(sp, &iv.Hash, c, waitChan, wire.BaseEncoding)
case wire.InvTypeWitnessBlock:
err = sp.server.pushBlockMsg(sp, &iv.Hash, c, waitChan, wire.WitnessEncoding)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
case wire.InvTypeBlock:
err = sp.server.pushBlockMsg(sp, &iv.Hash, c, waitChan, wire.BaseEncoding)
case wire.InvTypeFilteredWitnessBlock:
err = sp.server.pushMerkleBlockMsg(sp, &iv.Hash, c, waitChan, wire.WitnessEncoding)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
case wire.InvTypeFilteredBlock:
err = sp.server.pushMerkleBlockMsg(sp, &iv.Hash, c, waitChan, wire.BaseEncoding)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
default:
peerLog.Warnf("Unknown type in inventory request %d",
iv.Type)
continue
}
if err != nil {
notFound.AddInvVect(iv)
// When there is a failure fetching the final entry
// and the done channel was sent in due to there
// being no outstanding not found inventory, consume
// it here because there is now not found inventory
// that will use the channel momentarily.
if i == len(msg.InvList)-1 && c != nil {
<-c
}
}
numAdded++
waitChan = c
}
if len(notFound.InvList) != 0 {
sp.QueueMessage(notFound, doneChan)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
}
// Wait for messages to be sent. We can send quite a lot of data at this
// point and this will keep the peer busy for a decent amount of time.
// We don't process anything else by them in this time so that we
// have an idea of when we should hear back from them - else the idle
// timeout could fire when we were only half done sending the blocks.
if numAdded > 0 {
<-doneChan
}
}
// OnGetBlocks is invoked when a peer receives a getblocks bitcoin
// message.
func (sp *serverPeer) OnGetBlocks(_ *peer.Peer, msg *wire.MsgGetBlocks) {
// Find the most recent known block in the best chain based on the block
// locator and fetch all of the block hashes after it until either
// wire.MaxBlocksPerMsg have been fetched or the provided stop hash is
// encountered.
//
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// Use the block after the genesis block if no other blocks in the
// provided locator are known. This does mean the client will start
// over with the genesis block if unknown block locators are provided.
//
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// This mirrors the behavior in the reference implementation.
chain := sp.server.chain
hashList := chain.LocateBlocks(msg.BlockLocatorHashes, &msg.HashStop,
wire.MaxBlocksPerMsg)
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
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// Generate inventory message.
invMsg := wire.NewMsgInv()
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
for i := range hashList {
iv := wire.NewInvVect(wire.InvTypeBlock, &hashList[i])
invMsg.AddInvVect(iv)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
}
// Send the inventory message if there is anything to send.
if len(invMsg.InvList) > 0 {
invListLen := len(invMsg.InvList)
if invListLen == wire.MaxBlocksPerMsg {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// Intentionally use a copy of the final hash so there
// is not a reference into the inventory slice which
// would prevent the entire slice from being eligible
// for GC as soon as it's sent.
continueHash := invMsg.InvList[invListLen-1].Hash
sp.continueHash = &continueHash
}
sp.QueueMessage(invMsg, nil)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
}
}
// OnGetHeaders is invoked when a peer receives a getheaders bitcoin
// message.
func (sp *serverPeer) OnGetHeaders(_ *peer.Peer, msg *wire.MsgGetHeaders) {
// Ignore getheaders requests if not in sync.
if !sp.server.syncManager.IsCurrent() {
return
}
// Find the most recent known block in the best chain based on the block
// locator and fetch all of the headers after it until either
// wire.MaxBlockHeadersPerMsg have been fetched or the provided stop
// hash is encountered.
//
// Use the block after the genesis block if no other blocks in the
// provided locator are known. This does mean the client will start
// over with the genesis block if unknown block locators are provided.
//
// This mirrors the behavior in the reference implementation.
chain := sp.server.chain
headers := chain.LocateHeaders(msg.BlockLocatorHashes, &msg.HashStop)
// Send found headers to the requesting peer.
blockHeaders := make([]*wire.BlockHeader, len(headers))
for i := range headers {
blockHeaders[i] = &headers[i]
}
sp.QueueMessage(&wire.MsgHeaders{Headers: blockHeaders}, nil)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
}
// OnGetCFilters is invoked when a peer receives a getcfilters bitcoin message.
func (sp *serverPeer) OnGetCFilters(_ *peer.Peer, msg *wire.MsgGetCFilters) {
// Ignore getcfilters requests if not in sync.
if !sp.server.syncManager.IsCurrent() {
return
}
// We'll also ensure that the remote party is requesting a set of
// filters that we actually currently maintain.
switch msg.FilterType {
case wire.GCSFilterRegular:
break
default:
peerLog.Debug("Filter request for unknown filter: %v",
msg.FilterType)
return
}
hashes, err := sp.server.chain.HeightToHashRange(
int32(msg.StartHeight), &msg.StopHash, wire.MaxGetCFiltersReqRange,
)
if err != nil {
peerLog.Debugf("Invalid getcfilters request: %v", err)
return
}
// Create []*chainhash.Hash from []chainhash.Hash to pass to
// FiltersByBlockHashes.
hashPtrs := make([]*chainhash.Hash, len(hashes))
for i := range hashes {
hashPtrs[i] = &hashes[i]
}
filters, err := sp.server.cfIndex.FiltersByBlockHashes(
hashPtrs, msg.FilterType,
)
if err != nil {
peerLog.Errorf("Error retrieving cfilters: %v", err)
return
}
for i, filterBytes := range filters {
if len(filterBytes) == 0 {
peerLog.Warnf("Could not obtain cfilter for %v",
hashes[i])
return
}
filterMsg := wire.NewMsgCFilter(
msg.FilterType, &hashes[i], filterBytes,
)
sp.QueueMessage(filterMsg, nil)
}
}
// OnGetCFHeaders is invoked when a peer receives a getcfheader bitcoin message.
func (sp *serverPeer) OnGetCFHeaders(_ *peer.Peer, msg *wire.MsgGetCFHeaders) {
2017-02-02 11:42:44 +01:00
// Ignore getcfilterheader requests if not in sync.
if !sp.server.syncManager.IsCurrent() {
2017-02-02 11:42:44 +01:00
return
}
// We'll also ensure that the remote party is requesting a set of
// headers for filters that we actually currently maintain.
switch msg.FilterType {
case wire.GCSFilterRegular:
break
default:
peerLog.Debug("Filter request for unknown headers for "+
"filter: %v", msg.FilterType)
return
}
startHeight := int32(msg.StartHeight)
maxResults := wire.MaxCFHeadersPerMsg
// If StartHeight is positive, fetch the predecessor block hash so we
// can populate the PrevFilterHeader field.
if msg.StartHeight > 0 {
startHeight--
maxResults++
}
// Fetch the hashes from the block index.
hashList, err := sp.server.chain.HeightToHashRange(
startHeight, &msg.StopHash, maxResults,
)
if err != nil {
peerLog.Debugf("Invalid getcfheaders request: %v", err)
}
// This is possible if StartHeight is one greater that the height of
// StopHash, and we pull a valid range of hashes including the previous
// filter header.
if len(hashList) == 0 || (msg.StartHeight > 0 && len(hashList) == 1) {
peerLog.Debug("No results for getcfheaders request")
return
}
// Create []*chainhash.Hash from []chainhash.Hash to pass to
// FilterHeadersByBlockHashes.
hashPtrs := make([]*chainhash.Hash, len(hashList))
for i := range hashList {
hashPtrs[i] = &hashList[i]
}
// Fetch the raw filter hash bytes from the database for all blocks.
filterHashes, err := sp.server.cfIndex.FilterHashesByBlockHashes(
hashPtrs, msg.FilterType,
)
if err != nil {
peerLog.Errorf("Error retrieving cfilter hashes: %v", err)
return
}
// Generate cfheaders message and send it.
headersMsg := wire.NewMsgCFHeaders()
// Populate the PrevFilterHeader field.
if msg.StartHeight > 0 {
prevBlockHash := &hashList[0]
// Fetch the raw committed filter header bytes from the
// database.
headerBytes, err := sp.server.cfIndex.FilterHeaderByBlockHash(
prevBlockHash, msg.FilterType)
if err != nil {
peerLog.Errorf("Error retrieving CF header: %v", err)
return
}
if len(headerBytes) == 0 {
peerLog.Warnf("Could not obtain CF header for %v", prevBlockHash)
return
}
// Deserialize the hash into PrevFilterHeader.
err = headersMsg.PrevFilterHeader.SetBytes(headerBytes)
if err != nil {
peerLog.Warnf("Committed filter header deserialize "+
"failed: %v", err)
return
}
hashList = hashList[1:]
filterHashes = filterHashes[1:]
}
// Populate HeaderHashes.
for i, hashBytes := range filterHashes {
if len(hashBytes) == 0 {
peerLog.Warnf("Could not obtain CF hash for %v", hashList[i])
return
}
// Deserialize the hash.
filterHash, err := chainhash.NewHash(hashBytes)
if err != nil {
peerLog.Warnf("Committed filter hash deserialize "+
"failed: %v", err)
return
}
headersMsg.AddCFHash(filterHash)
2017-02-02 11:42:44 +01:00
}
headersMsg.FilterType = msg.FilterType
headersMsg.StopHash = msg.StopHash
sp.QueueMessage(headersMsg, nil)
2017-02-02 11:42:44 +01:00
}
// OnGetCFCheckpt is invoked when a peer receives a getcfcheckpt bitcoin message.
func (sp *serverPeer) OnGetCFCheckpt(_ *peer.Peer, msg *wire.MsgGetCFCheckpt) {
// Ignore getcfcheckpt requests if not in sync.
if !sp.server.syncManager.IsCurrent() {
return
}
// We'll also ensure that the remote party is requesting a set of
// checkpoints for filters that we actually currently maintain.
switch msg.FilterType {
case wire.GCSFilterRegular:
break
default:
peerLog.Debug("Filter request for unknown checkpoints for "+
"filter: %v", msg.FilterType)
return
}
// Now that we know the client is fetching a filter that we know of,
// we'll fetch the block hashes et each check point interval so we can
// compare against our cache, and create new check points if necessary.
blockHashes, err := sp.server.chain.IntervalBlockHashes(
&msg.StopHash, wire.CFCheckptInterval,
)
if err != nil {
peerLog.Debugf("Invalid getcfilters request: %v", err)
return
}
checkptMsg := wire.NewMsgCFCheckpt(
msg.FilterType, &msg.StopHash, len(blockHashes),
)
// Fetch the current existing cache so we can decide if we need to
// extend it or if its adequate as is.
sp.server.cfCheckptCachesMtx.RLock()
checkptCache := sp.server.cfCheckptCaches[msg.FilterType]
// If the set of block hashes is beyond the current size of the cache,
// then we'll expand the size of the cache and also retain the write
// lock.
var updateCache bool
if len(blockHashes) > len(checkptCache) {
// Now that we know we'll need to modify the size of the cache,
// we'll release the read lock and grab the write lock to
// possibly expand the cache size.
sp.server.cfCheckptCachesMtx.RUnlock()
sp.server.cfCheckptCachesMtx.Lock()
defer sp.server.cfCheckptCachesMtx.Unlock()
// Now that we have the write lock, we'll check again as it's
// possible that the cache has already been expanded.
checkptCache = sp.server.cfCheckptCaches[msg.FilterType]
// If we still need to expand the cache, then We'll mark that
// we need to update the cache for below and also expand the
// size of the cache in place.
if len(blockHashes) > len(checkptCache) {
updateCache = true
additionalLength := len(blockHashes) - len(checkptCache)
newEntries := make([]cfHeaderKV, additionalLength)
peerLog.Infof("Growing size of checkpoint cache from %v to %v "+
"block hashes", len(checkptCache), len(blockHashes))
checkptCache = append(
sp.server.cfCheckptCaches[msg.FilterType],
newEntries...,
)
}
} else {
// Otherwise, we'll hold onto the read lock for the remainder
// of this method.
defer sp.server.cfCheckptCachesMtx.RUnlock()
peerLog.Tracef("Serving stale cache of size %v",
len(checkptCache))
}
// Now that we know the cache is of an appropriate size, we'll iterate
// backwards until the find the block hash. We do this as it's possible
// a re-org has occurred so items in the db are now in the main china
// while the cache has been partially invalidated.
var forkIdx int
for forkIdx = len(blockHashes); forkIdx > 0; forkIdx-- {
if checkptCache[forkIdx-1].blockHash == blockHashes[forkIdx-1] {
break
}
}
// Now that we know the how much of the cache is relevant for this
// query, we'll populate our check point message with the cache as is.
// Shortly below, we'll populate the new elements of the cache.
for i := 0; i < forkIdx; i++ {
checkptMsg.AddCFHeader(&checkptCache[i].filterHeader)
}
// We'll now collect the set of hashes that are beyond our cache so we
// can look up the filter headers to populate the final cache.
blockHashPtrs := make([]*chainhash.Hash, 0, len(blockHashes)-forkIdx)
for i := forkIdx; i < len(blockHashes); i++ {
blockHashPtrs = append(blockHashPtrs, &blockHashes[i])
}
filterHeaders, err := sp.server.cfIndex.FilterHeadersByBlockHashes(
blockHashPtrs, msg.FilterType,
)
if err != nil {
peerLog.Errorf("Error retrieving cfilter headers: %v", err)
return
}
// Now that we have the full set of filter headers, we'll add them to
// the checkpoint message, and also update our cache in line.
for i, filterHeaderBytes := range filterHeaders {
if len(filterHeaderBytes) == 0 {
peerLog.Warnf("Could not obtain CF header for %v",
blockHashPtrs[i])
return
}
filterHeader, err := chainhash.NewHash(filterHeaderBytes)
if err != nil {
peerLog.Warnf("Committed filter header deserialize "+
"failed: %v", err)
return
}
checkptMsg.AddCFHeader(filterHeader)
// If the new main chain is longer than what's in the cache,
// then we'll override it beyond the fork point.
if updateCache {
checkptCache[forkIdx+i] = cfHeaderKV{
blockHash: blockHashes[forkIdx+i],
filterHeader: *filterHeader,
}
}
}
// Finally, we'll update the cache if we need to, and send the final
// message back to the requesting peer.
if updateCache {
sp.server.cfCheckptCaches[msg.FilterType] = checkptCache
}
sp.QueueMessage(checkptMsg, nil)
}
// enforceNodeBloomFlag disconnects the peer if the server is not configured to
// allow bloom filters. Additionally, if the peer has negotiated to a protocol
// version that is high enough to observe the bloom filter service support bit,
// it will be banned since it is intentionally violating the protocol.
func (sp *serverPeer) enforceNodeBloomFlag(cmd string) bool {
if sp.server.services&wire.SFNodeBloom != wire.SFNodeBloom {
// Ban the peer if the protocol version is high enough that the
// peer is knowingly violating the protocol and banning is
// enabled.
//
// NOTE: Even though the addBanScore function already examines
// whether or not banning is enabled, it is checked here as well
// to ensure the violation is logged and the peer is
// disconnected regardless.
if sp.ProtocolVersion() >= wire.BIP0111Version &&
!cfg.DisableBanning {
// Disconnect the peer regardless of whether it was
// banned.
sp.addBanScore(100, 0, cmd)
sp.Disconnect()
return false
}
// Disconnect the peer regardless of protocol version or banning
// state.
peerLog.Debugf("%s sent an unsupported %s request -- "+
"disconnecting", sp, cmd)
sp.Disconnect()
return false
}
return true
}
// OnFeeFilter is invoked when a peer receives a feefilter bitcoin message and
// is used by remote peers to request that no transactions which have a fee rate
// lower than provided value are inventoried to them. The peer will be
// disconnected if an invalid fee filter value is provided.
func (sp *serverPeer) OnFeeFilter(_ *peer.Peer, msg *wire.MsgFeeFilter) {
// Check that the passed minimum fee is a valid amount.
if msg.MinFee < 0 || msg.MinFee > btcutil.MaxSatoshi {
peerLog.Debugf("Peer %v sent an invalid feefilter '%v' -- "+
"disconnecting", sp, btcutil.Amount(msg.MinFee))
sp.Disconnect()
return
}
atomic.StoreInt64(&sp.feeFilter, msg.MinFee)
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// OnFilterAdd is invoked when a peer receives a filteradd bitcoin
// message and is used by remote peers to add data to an already loaded bloom
// filter. The peer will be disconnected if a filter is not loaded when this
// message is received or the server is not configured to allow bloom filters.
func (sp *serverPeer) OnFilterAdd(_ *peer.Peer, msg *wire.MsgFilterAdd) {
// Disconnect and/or ban depending on the node bloom services flag and
// negotiated protocol version.
if !sp.enforceNodeBloomFlag(msg.Command()) {
return
}
if !sp.filter.IsLoaded() {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
peerLog.Debugf("%s sent a filteradd request with no filter "+
"loaded -- disconnecting", sp)
sp.Disconnect()
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
return
}
sp.filter.Add(msg.Data)
}
// OnFilterClear is invoked when a peer receives a filterclear bitcoin
// message and is used by remote peers to clear an already loaded bloom filter.
// The peer will be disconnected if a filter is not loaded when this message is
// received or the server is not configured to allow bloom filters.
func (sp *serverPeer) OnFilterClear(_ *peer.Peer, msg *wire.MsgFilterClear) {
// Disconnect and/or ban depending on the node bloom services flag and
// negotiated protocol version.
if !sp.enforceNodeBloomFlag(msg.Command()) {
return
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
if !sp.filter.IsLoaded() {
peerLog.Debugf("%s sent a filterclear request with no "+
"filter loaded -- disconnecting", sp)
sp.Disconnect()
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
return
}
sp.filter.Unload()
}
// OnFilterLoad is invoked when a peer receives a filterload bitcoin
// message and it used to load a bloom filter that should be used for
// delivering merkle blocks and associated transactions that match the filter.
// The peer will be disconnected if the server is not configured to allow bloom
// filters.
func (sp *serverPeer) OnFilterLoad(_ *peer.Peer, msg *wire.MsgFilterLoad) {
// Disconnect and/or ban depending on the node bloom services flag and
// negotiated protocol version.
if !sp.enforceNodeBloomFlag(msg.Command()) {
return
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
sp.setDisableRelayTx(false)
sp.filter.Reload(msg)
}
// OnGetAddr is invoked when a peer receives a getaddr bitcoin message
// and is used to provide the peer with known addresses from the address
// manager.
func (sp *serverPeer) OnGetAddr(_ *peer.Peer, msg *wire.MsgGetAddr) {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// Don't return any addresses when running on the simulation test
// network. This helps prevent the network from becoming another
// public test network since it will not be able to learn about other
// peers that have not specifically been provided.
if cfg.SimNet {
return
}
// Do not accept getaddr requests from outbound peers. This reduces
// fingerprinting attacks.
if !sp.Inbound() {
peerLog.Debugf("Ignoring getaddr request from outbound peer "+
"%v", sp)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
return
}
// Only allow one getaddr request per connection to discourage
// address stamping of inv announcements.
if sp.sentAddrs {
peerLog.Debugf("Ignoring repeated getaddr request from peer "+
"%v", sp)
return
}
sp.sentAddrs = true
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// Get the current known addresses from the address manager.
addrCache := sp.server.addrManager.AddressCache()
// Push the addresses.
sp.pushAddrMsg(addrCache)
}
// OnAddr is invoked when a peer receives an addr bitcoin message and is
// used to notify the server about advertised addresses.
func (sp *serverPeer) OnAddr(_ *peer.Peer, msg *wire.MsgAddr) {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// Ignore addresses when running on the simulation test network. This
// helps prevent the network from becoming another public test network
// since it will not be able to learn about other peers that have not
// specifically been provided.
if cfg.SimNet {
return
}
// Ignore old style addresses which don't include a timestamp.
if sp.ProtocolVersion() < wire.NetAddressTimeVersion {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
return
}
// A message that has no addresses is invalid.
if len(msg.AddrList) == 0 {
peerLog.Errorf("Command [%s] from %s does not contain any addresses",
2018-05-16 05:06:16 +02:00
msg.Command(), sp.Peer)
sp.Disconnect()
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
return
}
for _, na := range msg.AddrList {
// Don't add more address if we're disconnecting.
if !sp.Connected() {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
return
}
// Set the timestamp to 5 days ago if it's more than 24 hours
// in the future so this address is one of the first to be
// removed when space is needed.
now := time.Now()
if na.Timestamp.After(now.Add(time.Minute * 10)) {
na.Timestamp = now.Add(-1 * time.Hour * 24 * 5)
}
// Add address to known addresses for this peer.
sp.addKnownAddresses([]*wire.NetAddress{na})
}
// Add addresses to server address manager. The address manager handles
// the details of things such as preventing duplicate addresses, max
// addresses, and last seen updates.
// XXX bitcoind gives a 2 hour time penalty here, do we want to do the
// same?
sp.server.addrManager.AddAddresses(msg.AddrList, sp.NA())
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
}
// OnRead is invoked when a peer receives a message and it is used to update
// the bytes received by the server.
func (sp *serverPeer) OnRead(_ *peer.Peer, bytesRead int, msg wire.Message, err error) {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
sp.server.AddBytesReceived(uint64(bytesRead))
}
// OnWrite is invoked when a peer sends a message and it is used to update
// the bytes sent by the server.
func (sp *serverPeer) OnWrite(_ *peer.Peer, bytesWritten int, msg wire.Message, err error) {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
sp.server.AddBytesSent(uint64(bytesWritten))
}
// OnNotFound is invoked when a peer sends a notfound message.
func (sp *serverPeer) OnNotFound(p *peer.Peer, msg *wire.MsgNotFound) {
if !sp.Connected() {
return
}
var numBlocks, numTxns uint32
for _, inv := range msg.InvList {
switch inv.Type {
case wire.InvTypeBlock:
numBlocks++
case wire.InvTypeWitnessBlock:
numBlocks++
case wire.InvTypeTx:
numTxns++
case wire.InvTypeWitnessTx:
numTxns++
default:
peerLog.Infof("Invalid inv type '%d' in NotFound message from %s. Disconnecting...", inv.Type, sp)
sp.Disconnect()
return
}
}
if numBlocks > 0 {
blockStr := pickNoun(uint64(numBlocks), "block", "blocks")
reason := fmt.Sprintf("%d %v not found on %s", numBlocks, blockStr, sp)
if sp.addBanScore(20, 0, reason) {
return // once they fail to return us five block requests they're gone for good
}
}
if numTxns > 0 {
// This is an expected situation if transactions in the mempool make it into a block before
// this node knows about said block. We don't want to ban them for that alone
peerLog.Debugf("%d transactions not found on %s", numTxns, sp)
if numBlocks+numTxns < wire.MaxInvPerMsg { // if our message is full then it is likely followed by another one that isn't
txStr := pickNoun(uint64(numTxns), "transaction", "transactions")
reason := fmt.Sprintf("%d %v not found on %s", numTxns, txStr, sp)
if sp.addBanScore(0, 20, reason) {
return // if they fail us five times in one minute, they're gone -- hitting them at new-block should be rare
}
}
}
sp.server.syncManager.QueueNotFound(msg, p)
}
// randomUint16Number returns a random uint16 in a specified input range. Note
// that the range is in zeroth ordering; if you pass it 1800, you will get
// values from 0 to 1800.
func randomUint16Number(max uint16) uint16 {
// In order to avoid modulo bias and ensure every possible outcome in
// [0, max) has equal probability, the random number must be sampled
// from a random source that has a range limited to a multiple of the
// modulus.
var randomNumber uint16
var limitRange = (math.MaxUint16 / max) * max
for {
binary.Read(rand.Reader, binary.LittleEndian, &randomNumber)
if randomNumber < limitRange {
return (randomNumber % max)
}
}
}
// AddRebroadcastInventory adds 'iv' to the list of inventories to be
// rebroadcasted at random intervals until they show up in a block.
func (s *server) AddRebroadcastInventory(iv *wire.InvVect, data interface{}) {
// Ignore if shutting down.
if atomic.LoadInt32(&s.shutdown) != 0 {
return
}
s.modifyRebroadcastInv <- broadcastInventoryAdd{invVect: iv, data: data}
}
// RemoveRebroadcastInventory removes 'iv' from the list of items to be
// rebroadcasted if present.
func (s *server) RemoveRebroadcastInventory(iv *wire.InvVect) {
// Ignore if shutting down.
if atomic.LoadInt32(&s.shutdown) != 0 {
return
}
s.modifyRebroadcastInv <- broadcastInventoryDel(iv)
}
// relayTransactions generates and relays inventory vectors for all of the
// passed transactions to all connected peers.
func (s *server) relayTransactions(txns []*mempool.TxDesc) {
for _, txD := range txns {
iv := wire.NewInvVect(wire.InvTypeTx, txD.Tx.Hash())
s.RelayInventory(iv, txD)
}
}
// AnnounceNewTransactions generates and relays inventory vectors and notifies
// both websocket and getblocktemplate long poll clients of the passed
// transactions. This function should be called whenever new transactions
// are added to the mempool.
func (s *server) AnnounceNewTransactions(txns []*mempool.TxDesc) {
// Generate and relay inventory vectors for all newly accepted
// transactions.
s.relayTransactions(txns)
// Notify both websocket and getblocktemplate long poll clients of all
// newly accepted transactions.
if s.rpcServer != nil {
s.rpcServer.NotifyNewTransactions(txns)
}
}
// Transaction has one confirmation on the main chain. Now we can mark it as no
// longer needing rebroadcasting.
func (s *server) TransactionConfirmed(tx *btcutil.Tx) {
// Rebroadcasting is only necessary when the RPC server is active.
if s.rpcServer == nil {
return
}
iv := wire.NewInvVect(wire.InvTypeTx, tx.Hash())
s.RemoveRebroadcastInventory(iv)
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// pushTxMsg sends a tx message for the provided transaction hash to the
// connected peer. An error is returned if the transaction hash is not known.
func (s *server) pushTxMsg(sp *serverPeer, hash *chainhash.Hash, doneChan chan<- struct{},
waitChan <-chan struct{}, encoding wire.MessageEncoding) error {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// Attempt to fetch the requested transaction from the pool. A
// call could be made to check for existence first, but simply trying
// to fetch a missing transaction results in the same behavior.
tx, err := s.txMemPool.FetchTransaction(hash)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
if err != nil {
peerLog.Tracef("Unable to fetch tx %v from transaction "+
"pool: %v", hash, err)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
if doneChan != nil {
doneChan <- struct{}{}
}
return err
}
// Once we have fetched data wait for any previous operation to finish.
if waitChan != nil {
<-waitChan
}
sp.QueueMessageWithEncoding(tx.MsgTx(), doneChan, encoding)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
return nil
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// pushBlockMsg sends a block message for the provided block hash to the
// connected peer. An error is returned if the block hash is not known.
func (s *server) pushBlockMsg(sp *serverPeer, hash *chainhash.Hash, doneChan chan<- struct{},
waitChan <-chan struct{}, encoding wire.MessageEncoding) 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
// Fetch the raw block bytes from the database.
var blockBytes []byte
err := sp.server.db.View(func(dbTx database.Tx) error {
var err error
blockBytes, err = dbTx.FetchBlock(hash)
return err
})
if err != nil {
peerLog.Tracef("Unable to fetch requested block hash %v: %v",
hash, err)
if doneChan != nil {
doneChan <- struct{}{}
}
return err
}
// Deserialize the block.
var msgBlock wire.MsgBlock
err = msgBlock.Deserialize(bytes.NewReader(blockBytes))
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
if err != nil {
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
peerLog.Tracef("Unable to deserialize requested block hash "+
"%v: %v", hash, err)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
if doneChan != nil {
doneChan <- struct{}{}
}
return err
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// Once we have fetched data wait for any previous operation to finish.
if waitChan != nil {
<-waitChan
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// We only send the channel for this message if we aren't sending
// an inv straight after.
var dc chan<- struct{}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
continueHash := sp.continueHash
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
sendInv := continueHash != nil && continueHash.IsEqual(hash)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
if !sendInv {
dc = doneChan
}
sp.QueueMessageWithEncoding(&msgBlock, dc, encoding)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// When the peer requests the final block that was advertised in
// response to a getblocks message which requested more blocks than
// would fit into a single message, send it a new inventory message
// to trigger it to issue another getblocks message for the next
// batch of inventory.
if sendInv {
best := sp.server.chain.BestSnapshot()
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
invMsg := wire.NewMsgInvSizeHint(1)
iv := wire.NewInvVect(wire.InvTypeBlock, &best.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
invMsg.AddInvVect(iv)
sp.QueueMessage(invMsg, doneChan)
sp.continueHash = nil
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
}
return nil
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// pushMerkleBlockMsg sends a merkleblock message for the provided block hash to
// the connected peer. Since a merkle block requires the peer to have a filter
// loaded, this call will simply be ignored if there is no filter loaded. An
// error is returned if the block hash is not known.
func (s *server) pushMerkleBlockMsg(sp *serverPeer, hash *chainhash.Hash,
doneChan chan<- struct{}, waitChan <-chan struct{}, encoding wire.MessageEncoding) error {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// Do not send a response if the peer doesn't have a filter loaded.
if !sp.filter.IsLoaded() {
if doneChan != nil {
doneChan <- struct{}{}
}
return nil
}
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 the raw block bytes from the database.
blk, err := sp.server.chain.BlockByHash(hash)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
if err != nil {
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
peerLog.Tracef("Unable to fetch requested block hash %v: %v",
hash, err)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
if doneChan != nil {
doneChan <- struct{}{}
}
return err
}
// Generate a merkle block by filtering the requested block according
// to the filter for the peer.
merkle, matchedTxIndices := bloom.NewMerkleBlock(blk, sp.filter)
// Once we have fetched data wait for any previous operation to finish.
if waitChan != nil {
<-waitChan
}
// Send the merkleblock. Only send the done channel with this message
// if no transactions will be sent afterwards.
var dc chan<- struct{}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
if len(matchedTxIndices) == 0 {
dc = doneChan
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
sp.QueueMessage(merkle, dc)
// Finally, send any matched transactions.
blkTransactions := blk.MsgBlock().Transactions
for i, txIndex := range matchedTxIndices {
// Only send the done channel on the final transaction.
var dc chan<- struct{}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
if i == len(matchedTxIndices)-1 {
dc = doneChan
}
if txIndex < uint32(len(blkTransactions)) {
sp.QueueMessageWithEncoding(blkTransactions[txIndex], dc,
encoding)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
}
}
return nil
}
// handleUpdatePeerHeight updates the heights of all peers who were known to
// announce a block we recently accepted.
func (s *server) handleUpdatePeerHeights(state *peerState, umsg updatePeerHeightsMsg) {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
state.forAllPeers(func(sp *serverPeer) {
// The origin peer should already have the updated height.
if sp.Peer == umsg.originPeer {
return
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// This is a pointer to the underlying memory which doesn't
// change.
latestBlkHash := sp.LastAnnouncedBlock()
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// Skip this peer if it hasn't recently announced any new blocks.
if latestBlkHash == nil {
return
}
// If the peer has recently announced a block, and this block
// matches our newly accepted block, then update their block
// height.
if *latestBlkHash == *umsg.newHash {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
sp.UpdateLastBlockHeight(umsg.newHeight)
sp.UpdateLastAnnouncedBlock(nil)
}
})
}
2013-08-06 23:55:22 +02:00
// handleAddPeerMsg deals with adding new peers. It is invoked from the
// peerHandler goroutine.
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
func (s *server) handleAddPeerMsg(state *peerState, sp *serverPeer) bool {
if sp == nil || !sp.Connected() {
return false
}
// Disconnect peers with unwanted user agents.
if sp.HasUndesiredUserAgent(s.agentBlacklist, s.agentWhitelist) {
sp.Disconnect()
return false
}
2013-08-06 23:55:22 +02:00
// Ignore new peers if we're shutting down.
if atomic.LoadInt32(&s.shutdown) != 0 {
srvrLog.Infof("New peer %s ignored - server is shutting down", sp)
sp.Disconnect()
return false
2013-08-06 23:55:22 +02:00
}
// Disconnect banned peers.
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
host, _, err := net.SplitHostPort(sp.Addr())
2013-08-06 23:55:22 +02:00
if err != nil {
srvrLog.Debugf("can't split hostport %v", err)
sp.Disconnect()
return false
2013-08-06 23:55:22 +02:00
}
if banEnd, ok := state.banned[host]; ok {
2013-08-06 23:55:22 +02:00
if time.Now().Before(banEnd) {
srvrLog.Debugf("Peer %s is banned for another %v - disconnecting",
host, time.Until(banEnd))
sp.Disconnect()
return false
2013-08-06 23:55:22 +02:00
}
srvrLog.Infof("Peer %s is no longer banned", host)
delete(state.banned, host)
2013-08-06 23:55:22 +02:00
}
// TODO: Check for max peers from a single IP.
// Limit max number of total peers.
if state.Count() >= cfg.MaxPeers {
srvrLog.Infof("Max peers reached [%d] - disconnecting peer %s",
cfg.MaxPeers, sp)
sp.Disconnect()
2016-11-16 00:43:43 +01:00
// TODO: how to handle permanent peers here?
// they should be rescheduled.
return false
2013-08-06 23:55:22 +02:00
}
// Add the new peer and start it.
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
srvrLog.Debugf("New peer %s", sp)
if sp.Inbound() {
state.inboundPeers[sp.ID()] = sp
} else {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
state.outboundGroups[addrmgr.GroupKey(sp.NA())]++
if sp.persistent {
state.persistentPeers[sp.ID()] = sp
} else {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
state.outboundPeers[sp.ID()] = sp
}
}
// Update the address' last seen time if the peer has acknowledged
// our version and has sent us its version as well.
if sp.VerAckReceived() && sp.VersionKnown() && sp.NA() != nil {
s.addrManager.Connected(sp.NA())
}
// Signal the sync manager this peer is a new sync candidate.
s.syncManager.NewPeer(sp.Peer)
// Update the address manager and request known addresses from the
// remote peer for outbound connections. This is skipped when running on
// the simulation test network since it is only intended to connect to
// specified peers and actively avoids advertising and connecting to
// discovered peers.
if !cfg.SimNet && !sp.Inbound() {
// Advertise the local address when the server accepts incoming
// connections and it believes itself to be close to the best
// known tip.
if !cfg.DisableListen && s.syncManager.IsCurrent() {
// Get address that best matches.
lna := s.addrManager.GetBestLocalAddress(sp.NA())
if addrmgr.IsRoutable(lna) {
// Filter addresses the peer already knows about.
addresses := []*wire.NetAddress{lna}
sp.pushAddrMsg(addresses)
}
}
// Request known addresses if the server address manager needs
// more and the peer has a protocol version new enough to
// include a timestamp with addresses.
hasTimestamp := sp.ProtocolVersion() >= wire.NetAddressTimeVersion
if s.addrManager.NeedMoreAddresses() && hasTimestamp {
sp.QueueMessage(wire.NewMsgGetAddr(), nil)
}
// Mark the address as a known good address.
s.addrManager.Good(sp.NA())
}
return true
2013-08-06 23:55:22 +02:00
}
// handleDonePeerMsg deals with peers that have signalled they are done. It is
// invoked from the peerHandler goroutine.
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
func (s *server) handleDonePeerMsg(state *peerState, sp *serverPeer) {
var list map[int32]*serverPeer
if sp.persistent {
list = state.persistentPeers
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
} else if sp.Inbound() {
list = state.inboundPeers
} else {
list = state.outboundPeers
}
// Regardless of whether the peer was found in our list, we'll inform
// our connection manager about the disconnection. This can happen if we
// process a peer's `done` message before its `add`.
if !sp.Inbound() {
if sp.persistent {
s.connManager.Disconnect(sp.connReq.ID())
} else {
s.connManager.Remove(sp.connReq.ID())
go s.connManager.NewConnReq()
}
}
if _, ok := list[sp.ID()]; ok {
if !sp.Inbound() && sp.VersionKnown() {
state.outboundGroups[addrmgr.GroupKey(sp.NA())]--
}
delete(list, sp.ID())
srvrLog.Debugf("Removed peer %s", sp)
return
2013-08-06 23:55:22 +02:00
}
}
// handleBanPeerMsg deals with banning peers. It is invoked from the
// peerHandler goroutine.
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
func (s *server) handleBanPeerMsg(state *peerState, sp *serverPeer) {
host, _, err := net.SplitHostPort(sp.Addr())
2013-08-06 23:55:22 +02:00
if err != nil {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
srvrLog.Debugf("can't split ban peer %s %v", sp.Addr(), err)
2013-08-06 23:55:22 +02:00
return
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
direction := directionString(sp.Inbound())
srvrLog.Infof("Banned peer %s (%s) for %v", host, direction,
2013-08-06 23:55:22 +02:00
cfg.BanDuration)
state.banned[host] = time.Now().Add(cfg.BanDuration)
2013-08-06 23:55:22 +02:00
}
// handleRelayInvMsg deals with relaying inventory to peers that are not already
// known to have it. It is invoked from the peerHandler goroutine.
func (s *server) handleRelayInvMsg(state *peerState, msg relayMsg) {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
state.forAllPeers(func(sp *serverPeer) {
if !sp.Connected() {
return
}
// If the inventory is a block and the peer prefers headers,
// generate and send a headers message instead of an inventory
// message.
if msg.invVect.Type == wire.InvTypeBlock && sp.WantsHeaders() {
blockHeader, ok := msg.data.(wire.BlockHeader)
if !ok {
peerLog.Warnf("Underlying data for headers" +
" is not a block header")
return
}
msgHeaders := wire.NewMsgHeaders()
if err := msgHeaders.AddBlockHeader(&blockHeader); err != nil {
peerLog.Errorf("Failed to add block"+
" header: %v", err)
return
}
sp.QueueMessage(msgHeaders, nil)
return
}
if msg.invVect.Type == wire.InvTypeTx {
// Don't relay the transaction to the peer when it has
// transaction relaying disabled.
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
if sp.relayTxDisabled() {
return
}
txD, ok := msg.data.(*mempool.TxDesc)
if !ok {
2017-01-16 17:08:37 +01:00
peerLog.Warnf("Underlying data for tx inv "+
"relay is not a *mempool.TxDesc: %T",
msg.data)
return
}
// Don't relay the transaction if the transaction fee-per-kb
// is less than the peer's feefilter.
feeFilter := atomic.LoadInt64(&sp.feeFilter)
if feeFilter > 0 && txD.FeePerKB < feeFilter {
return
}
// Don't relay the transaction if there is a bloom
// filter loaded and the transaction doesn't match it.
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
if sp.filter.IsLoaded() {
if !sp.filter.MatchTxAndUpdate(txD.Tx) {
return
}
}
}
// Queue the inventory to be relayed with the next batch.
// It will be ignored if the peer is already known to
// have the inventory.
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
sp.QueueInventory(msg.invVect)
})
}
2013-08-06 23:55:22 +02:00
// handleBroadcastMsg deals with broadcasting messages to peers. It is invoked
// from the peerHandler goroutine.
func (s *server) handleBroadcastMsg(state *peerState, bmsg *broadcastMsg) {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
state.forAllPeers(func(sp *serverPeer) {
if !sp.Connected() {
return
}
for _, ep := range bmsg.excludePeers {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
if sp == ep {
return
2013-08-06 23:55:22 +02:00
}
}
sp.QueueMessage(bmsg.message, nil)
})
2013-08-06 23:55:22 +02:00
}
type getConnCountMsg struct {
reply chan int32
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
type getPeersMsg struct {
reply chan []*serverPeer
}
type getOutboundGroup struct {
key string
reply chan int
}
type getAddedNodesMsg struct {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
reply chan []*serverPeer
}
type disconnectNodeMsg struct {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
cmp func(*serverPeer) bool
reply chan error
}
type connectNodeMsg struct {
addr string
permanent bool
reply chan error
}
type removeNodeMsg struct {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
cmp func(*serverPeer) bool
reply chan error
}
2013-10-29 21:29:23 +01:00
// handleQuery is the central handler for all queries and commands from other
// goroutines related to peer state.
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
func (s *server) handleQuery(state *peerState, querymsg interface{}) {
switch msg := querymsg.(type) {
case getConnCountMsg:
nconnected := int32(0)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
state.forAllPeers(func(sp *serverPeer) {
if sp.Connected() {
nconnected++
}
})
msg.reply <- nconnected
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
case getPeersMsg:
peers := make([]*serverPeer, 0, state.Count())
state.forAllPeers(func(sp *serverPeer) {
if !sp.Connected() {
return
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
peers = append(peers, sp)
})
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
msg.reply <- peers
case connectNodeMsg:
2016-11-16 00:43:43 +01:00
// TODO: duplicate oneshots?
// Limit max number of total peers.
if state.Count() >= cfg.MaxPeers {
msg.reply <- errors.New("max peers reached")
return
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
for _, peer := range state.persistentPeers {
if peer.Addr() == msg.addr {
if msg.permanent {
msg.reply <- errors.New("peer already connected")
} else {
msg.reply <- errors.New("peer exists as a permanent peer")
}
return
}
}
netAddr, err := addrStringToNetAddr(msg.addr)
if err != nil {
msg.reply <- err
return
}
2016-11-16 00:43:43 +01:00
// TODO: if too many, nuke a non-perm peer.
go s.connManager.Connect(&connmgr.ConnReq{
Addr: netAddr,
Permanent: msg.permanent,
})
msg.reply <- nil
case removeNodeMsg:
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
found := disconnectPeer(state.persistentPeers, msg.cmp, func(sp *serverPeer) {
// Keep group counts ok since we remove from
// the list now.
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
state.outboundGroups[addrmgr.GroupKey(sp.NA())]--
})
if found {
msg.reply <- nil
} else {
msg.reply <- errors.New("peer not found")
}
case getOutboundGroup:
count, ok := state.outboundGroups[msg.key]
if ok {
msg.reply <- count
} else {
msg.reply <- 0
}
// Request a list of the persistent (added) peers.
case getAddedNodesMsg:
// Respond with a slice of the relevant peers.
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
peers := make([]*serverPeer, 0, len(state.persistentPeers))
for _, sp := range state.persistentPeers {
peers = append(peers, sp)
}
msg.reply <- peers
case disconnectNodeMsg:
// Check inbound peers. We pass a nil callback since we don't
// require any additional actions on disconnect for inbound peers.
found := disconnectPeer(state.inboundPeers, msg.cmp, nil)
if found {
msg.reply <- nil
return
}
// Check outbound peers.
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
found = disconnectPeer(state.outboundPeers, msg.cmp, func(sp *serverPeer) {
// Keep group counts ok since we remove from
// the list now.
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
state.outboundGroups[addrmgr.GroupKey(sp.NA())]--
})
if found {
// If there are multiple outbound connections to the same
// ip:port, continue disconnecting them all until no such
// peers are found.
for found {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
found = disconnectPeer(state.outboundPeers, msg.cmp, func(sp *serverPeer) {
state.outboundGroups[addrmgr.GroupKey(sp.NA())]--
})
}
msg.reply <- nil
return
}
msg.reply <- errors.New("peer not found")
}
}
// disconnectPeer attempts to drop the connection of a targeted peer in the
// passed peer list. Targets are identified via usage of the passed
// `compareFunc`, which should return `true` if the passed peer is the target
// peer. This function returns true on success and false if the peer is unable
// to be located. If the peer is found, and the passed callback: `whenFound'
// isn't nil, we call it with the peer as the argument before it is removed
// from the peerList, and is disconnected from the server.
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
func disconnectPeer(peerList map[int32]*serverPeer, compareFunc func(*serverPeer) bool, whenFound func(*serverPeer)) bool {
for addr, peer := range peerList {
if compareFunc(peer) {
if whenFound != nil {
whenFound(peer)
}
// This is ok because we are not continuing
// to iterate so won't corrupt the loop.
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
delete(peerList, addr)
peer.Disconnect()
return true
}
}
return false
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// newPeerConfig returns the configuration for the given serverPeer.
func newPeerConfig(sp *serverPeer) *peer.Config {
return &peer.Config{
Listeners: peer.MessageListeners{
OnVersion: sp.OnVersion,
OnVerAck: sp.OnVerAck,
OnMemPool: sp.OnMemPool,
OnTx: sp.OnTx,
OnBlock: sp.OnBlock,
OnInv: sp.OnInv,
OnHeaders: sp.OnHeaders,
OnGetData: sp.OnGetData,
OnGetBlocks: sp.OnGetBlocks,
OnGetHeaders: sp.OnGetHeaders,
OnGetCFilters: sp.OnGetCFilters,
OnGetCFHeaders: sp.OnGetCFHeaders,
OnGetCFCheckpt: sp.OnGetCFCheckpt,
OnFeeFilter: sp.OnFeeFilter,
OnFilterAdd: sp.OnFilterAdd,
OnFilterClear: sp.OnFilterClear,
OnFilterLoad: sp.OnFilterLoad,
OnGetAddr: sp.OnGetAddr,
OnAddr: sp.OnAddr,
OnRead: sp.OnRead,
OnWrite: sp.OnWrite,
OnNotFound: sp.OnNotFound,
2015-11-10 19:20:29 +01:00
// Note: The reference client currently bans peers that send alerts
// not signed with its key. We could verify against their key, but
// since the reference client is currently unwilling to support
// other implementations' alert messages, we will not relay theirs.
2015-11-10 19:20:29 +01:00
OnAlert: nil,
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
},
NewestBlock: sp.newestBlock,
HostToNetAddress: sp.server.addrManager.HostToNetAddress,
Proxy: cfg.Proxy,
UserAgentName: userAgentName,
UserAgentVersion: userAgentVersion,
UserAgentComments: cfg.UserAgentComments,
ChainParams: sp.server.chainParams,
Services: sp.server.services,
DisableRelayTx: cfg.BlocksOnly,
ProtocolVersion: peer.MaxProtocolVersion,
TrickleInterval: cfg.TrickleInterval,
DisableStallHandler: cfg.DisableStallHandler,
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
}
}
// inboundPeerConnected is invoked by the connection manager when a new inbound
// connection is established. It initializes a new inbound server peer
// instance, associates it with the connection, and starts a goroutine to wait
// for disconnection.
func (s *server) inboundPeerConnected(conn net.Conn) {
sp := newServerPeer(s, false)
2017-08-31 15:59:43 +02:00
sp.isWhitelisted = isWhitelisted(conn.RemoteAddr())
sp.Peer = peer.NewInboundPeer(newPeerConfig(sp))
sp.AssociateConnection(conn)
go s.peerDoneHandler(sp)
2013-08-06 23:55:22 +02:00
}
// outboundPeerConnected is invoked by the connection manager when a new
// outbound connection is established. It initializes a new outbound server
// peer instance, associates it with the relevant state such as the connection
// request instance and the connection itself, and finally notifies the address
// manager of the attempt.
func (s *server) outboundPeerConnected(c *connmgr.ConnReq, conn net.Conn) {
sp := newServerPeer(s, c.Permanent)
p, err := peer.NewOutboundPeer(newPeerConfig(sp), c.Addr.String())
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
if err != nil {
srvrLog.Debugf("Cannot create outbound peer %s: %v", c.Addr, err)
if c.Permanent {
s.connManager.Disconnect(c.ID())
} else {
s.connManager.Remove(c.ID())
go s.connManager.NewConnReq()
}
return
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
}
sp.Peer = p
sp.connReq = c
2017-08-31 15:59:43 +02:00
sp.isWhitelisted = isWhitelisted(conn.RemoteAddr())
sp.AssociateConnection(conn)
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
go s.peerDoneHandler(sp)
}
// peerDoneHandler handles peer disconnects by notifiying the server that it's
// done along with other performing other desirable cleanup.
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
func (s *server) peerDoneHandler(sp *serverPeer) {
sp.WaitForDisconnect()
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
s.donePeers <- sp
// Only tell sync manager we are gone if we ever told it we existed.
if sp.VerAckReceived() {
s.syncManager.DonePeer(sp.Peer)
// Evict any remaining orphans that were sent by the peer.
numEvicted := s.txMemPool.RemoveOrphansByTag(mempool.Tag(sp.ID()))
if numEvicted > 0 {
txmpLog.Debugf("Evicted %d %s from peer %v (id %d)",
numEvicted, pickNoun(numEvicted, "orphan",
"orphans"), sp, sp.ID())
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
}
close(sp.quit)
}
2013-08-06 23:55:22 +02:00
// peerHandler is used to handle peer operations such as adding and removing
// peers to and from the server, banning peers, and broadcasting messages to
2014-09-08 21:19:47 +02:00
// peers. It must be run in a goroutine.
2013-08-06 23:55:22 +02:00
func (s *server) peerHandler() {
// Start the address manager and sync manager, both of which are needed
// by peers. This is done here since their lifecycle is closely tied
// to this handler and rather than adding more channels to sychronize
// things, it's easier and slightly faster to simply start and stop them
// in this handler.
s.addrManager.Start()
s.syncManager.Start()
srvrLog.Tracef("Starting peer handler")
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
state := &peerState{
inboundPeers: make(map[int32]*serverPeer),
persistentPeers: make(map[int32]*serverPeer),
outboundPeers: make(map[int32]*serverPeer),
banned: make(map[string]time.Time),
outboundGroups: make(map[string]int),
}
if !cfg.DisableDNSSeed {
// Add peers discovered through DNS to the address manager.
connmgr.SeedFromDNS(activeNetParams.Params, defaultRequiredServices,
btcdLookup, func(addrs []*wire.NetAddress) {
// Bitcoind uses a lookup of the dns seeder here. This
// is rather strange since the values looked up by the
// DNS seed lookups will vary quite a lot.
// to replicate this behaviour we put all addresses as
// having come from the first one.
s.addrManager.AddAddresses(addrs, addrs[0])
})
}
go s.connManager.Start()
out:
for {
2013-08-06 23:55:22 +02:00
select {
// New peers connected to the server.
case p := <-s.newPeers:
s.handleAddPeerMsg(state, p)
2013-08-06 23:55:22 +02:00
// Disconnected peers.
case p := <-s.donePeers:
s.handleDonePeerMsg(state, p)
2013-08-06 23:55:22 +02:00
// Block accepted in mainchain or orphan, update peer height.
case umsg := <-s.peerHeightsUpdate:
s.handleUpdatePeerHeights(state, umsg)
2013-08-06 23:55:22 +02:00
// Peer to ban.
case p := <-s.banPeers:
s.handleBanPeerMsg(state, p)
2013-08-06 23:55:22 +02:00
// New inventory to potentially be relayed to other peers.
case invMsg := <-s.relayInv:
s.handleRelayInvMsg(state, invMsg)
2013-08-06 23:55:22 +02:00
// Message to broadcast to all connected peers except those
// which are excluded by the message.
case bmsg := <-s.broadcast:
s.handleBroadcastMsg(state, &bmsg)
2013-08-06 23:55:22 +02:00
case qmsg := <-s.query:
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
s.handleQuery(state, qmsg)
2013-08-06 23:55:22 +02:00
case <-s.quit:
// Disconnect all peers on server shutdown.
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
state.forAllPeers(func(sp *serverPeer) {
srvrLog.Tracef("Shutdown peer %s", sp)
sp.Disconnect()
})
break out
2013-08-06 23:55:22 +02:00
}
}
s.connManager.Stop()
s.syncManager.Stop()
s.addrManager.Stop()
// Drain channels before exiting so nothing is left waiting around
// to send.
cleanup:
for {
select {
case <-s.newPeers:
case <-s.donePeers:
case <-s.peerHeightsUpdate:
case <-s.relayInv:
case <-s.broadcast:
case <-s.query:
default:
break cleanup
}
}
2013-08-06 23:55:22 +02:00
s.wg.Done()
srvrLog.Tracef("Peer handler done")
2013-08-06 23:55:22 +02:00
}
// AddPeer adds a new peer that has already been connected to the server.
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
func (s *server) AddPeer(sp *serverPeer) {
s.newPeers <- sp
2013-08-06 23:55:22 +02:00
}
// BanPeer bans a peer that has already been connected to the server by ip.
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
func (s *server) BanPeer(sp *serverPeer) {
s.banPeers <- sp
2013-08-06 23:55:22 +02:00
}
// RelayInventory relays the passed inventory vector to all connected peers
// that are not already known to have it.
func (s *server) RelayInventory(invVect *wire.InvVect, data interface{}) {
s.relayInv <- relayMsg{invVect: invVect, data: data}
}
2013-08-06 23:55:22 +02:00
// BroadcastMessage sends msg to all peers currently connected to the server
// except those in the passed peers to exclude.
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
func (s *server) BroadcastMessage(msg wire.Message, exclPeers ...*serverPeer) {
2013-08-06 23:55:22 +02:00
// XXX: Need to determine if this is an alert that has already been
// broadcast and refrain from broadcasting again.
bmsg := broadcastMsg{message: msg, excludePeers: exclPeers}
s.broadcast <- bmsg
}
2013-10-29 21:29:23 +01:00
// ConnectedCount returns the number of currently connected peers.
func (s *server) ConnectedCount() int32 {
replyChan := make(chan int32)
s.query <- getConnCountMsg{reply: replyChan}
return <-replyChan
}
// OutboundGroupCount returns the number of peers connected to the given
// outbound group key.
func (s *server) OutboundGroupCount(key string) int {
replyChan := make(chan int)
s.query <- getOutboundGroup{key: key, reply: replyChan}
return <-replyChan
}
// AddBytesSent adds the passed number of bytes to the total bytes sent counter
// for the server. It is safe for concurrent access.
func (s *server) AddBytesSent(bytesSent uint64) {
atomic.AddUint64(&s.bytesSent, bytesSent)
}
// AddBytesReceived adds the passed number of bytes to the total bytes received
// counter for the server. It is safe for concurrent access.
func (s *server) AddBytesReceived(bytesReceived uint64) {
atomic.AddUint64(&s.bytesReceived, bytesReceived)
}
// NetTotals returns the sum of all bytes received and sent across the network
// for all peers. It is safe for concurrent access.
func (s *server) NetTotals() (uint64, uint64) {
return atomic.LoadUint64(&s.bytesReceived),
atomic.LoadUint64(&s.bytesSent)
}
// UpdatePeerHeights updates the heights of all peers who have have announced
// the latest connected main chain block, or a recognized orphan. These height
// updates allow us to dynamically refresh peer heights, ensuring sync peer
// selection has access to the latest block heights for each peer.
func (s *server) UpdatePeerHeights(latestBlkHash *chainhash.Hash, latestHeight int32, updateSource *peer.Peer) {
s.peerHeightsUpdate <- updatePeerHeightsMsg{
newHash: latestBlkHash,
newHeight: latestHeight,
originPeer: updateSource,
}
}
// rebroadcastHandler keeps track of user submitted inventories that we have
// sent out but have not yet made it into a block. We periodically rebroadcast
// them in case our peers restarted or otherwise lost track of them.
func (s *server) rebroadcastHandler() {
// Wait 5 min before first tx rebroadcast.
timer := time.NewTimer(5 * time.Minute)
pendingInvs := make(map[wire.InvVect]interface{})
out:
for {
select {
case riv := <-s.modifyRebroadcastInv:
switch msg := riv.(type) {
// Incoming InvVects are added to our map of RPC txs.
case broadcastInventoryAdd:
pendingInvs[*msg.invVect] = msg.data
// When an InvVect has been added to a block, we can
// now remove it, if it was present.
case broadcastInventoryDel:
2020-05-13 14:58:39 +02:00
delete(pendingInvs, *msg)
}
case <-timer.C:
// Any inventory we have has not made it into a block
// yet. We periodically resubmit them until they have.
for iv, data := range pendingInvs {
ivCopy := iv
s.RelayInventory(&ivCopy, data)
}
// Process at a random time up to 30mins (in seconds)
// in the future.
timer.Reset(time.Second *
time.Duration(randomUint16Number(1800)))
case <-s.quit:
break out
}
}
timer.Stop()
// Drain channels before exiting so nothing is left waiting around
// to send.
cleanup:
for {
select {
case <-s.modifyRebroadcastInv:
default:
break cleanup
}
}
s.wg.Done()
}
2013-08-06 23:55:22 +02:00
// Start begins accepting connections from peers.
func (s *server) Start() {
// Already started?
if atomic.AddInt32(&s.started, 1) != 1 {
2013-08-06 23:55:22 +02:00
return
}
srvrLog.Trace("Starting server")
// Start the peer handler which in turn starts the address and block
// managers.
s.wg.Add(1)
go s.peerHandler()
if s.nat != nil {
s.wg.Add(1)
go s.upnpUpdateThread()
}
if !cfg.DisableRPC {
s.wg.Add(1)
// Start the rebroadcastHandler, which ensures user tx received by
// the RPC server are rebroadcast until being included in a block.
go s.rebroadcastHandler()
2013-08-06 23:55:22 +02:00
s.rpcServer.Start()
}
Implement a built-in concurrent CPU miner. This commit implements a built-in concurrent CPU miner that can be enabled with the combination of the --generate and --miningaddr options. The --blockminsize, --blockmaxsize, and --blockprioritysize configuration options wich already existed prior to this commit control the block template generation and hence affect blocks mined via the new CPU miner. The following is a quick overview of the changes and design: - Starting btcd with --generate and no addresses specified via --miningaddr will give an error and exit immediately - Makes use of multiple worker goroutines which independently create block templates, solve them, and submit the solved blocks - The default number of worker threads are based on the number of processor cores in the system and can be dynamically changed at run-time - There is a separate speed monitor goroutine used to collate periodic updates from the workers to calculate overall hashing speed - The current mining state, number of workers, and hashes per second can be queried - Updated sample-btcd.conf file has been updated to include the coin generation (mining) settings - Updated doc.go for the new command line options In addition the old --getworkkey option is now deprecated in favor of the new --miningaddr option. This was changed for a few reasons: - There is no reason to have a separate list of keys for getwork and CPU mining - getwork is deprecated and will be going away in the future so that means the --getworkkey flag will also be going away - Having the work 'key' in the option can be confused with wanting a private key while --miningaddr make it a little more clear it is an address that is required Closes #137. Reviewed by @jrick.
2014-06-12 03:09:38 +02:00
// Start the CPU miner if generation is enabled.
if cfg.Generate {
s.cpuMiner.Start()
}
2013-08-06 23:55:22 +02:00
}
// Stop gracefully shuts down the server by stopping and disconnecting all
// peers and the main listener.
func (s *server) Stop() error {
// Make sure this only happens once.
if atomic.AddInt32(&s.shutdown, 1) != 1 {
srvrLog.Infof("Server is already in the process of shutting down")
2013-08-06 23:55:22 +02:00
return nil
}
srvrLog.Warnf("Server shutting down")
Implement a built-in concurrent CPU miner. This commit implements a built-in concurrent CPU miner that can be enabled with the combination of the --generate and --miningaddr options. The --blockminsize, --blockmaxsize, and --blockprioritysize configuration options wich already existed prior to this commit control the block template generation and hence affect blocks mined via the new CPU miner. The following is a quick overview of the changes and design: - Starting btcd with --generate and no addresses specified via --miningaddr will give an error and exit immediately - Makes use of multiple worker goroutines which independently create block templates, solve them, and submit the solved blocks - The default number of worker threads are based on the number of processor cores in the system and can be dynamically changed at run-time - There is a separate speed monitor goroutine used to collate periodic updates from the workers to calculate overall hashing speed - The current mining state, number of workers, and hashes per second can be queried - Updated sample-btcd.conf file has been updated to include the coin generation (mining) settings - Updated doc.go for the new command line options In addition the old --getworkkey option is now deprecated in favor of the new --miningaddr option. This was changed for a few reasons: - There is no reason to have a separate list of keys for getwork and CPU mining - getwork is deprecated and will be going away in the future so that means the --getworkkey flag will also be going away - Having the work 'key' in the option can be confused with wanting a private key while --miningaddr make it a little more clear it is an address that is required Closes #137. Reviewed by @jrick.
2014-06-12 03:09:38 +02:00
// Stop the CPU miner if needed
s.cpuMiner.Stop()
// Shutdown the RPC server if it's not disabled.
if !cfg.DisableRPC {
s.rpcServer.Stop()
}
// Save fee estimator state in the database.
s.db.Update(func(tx database.Tx) error {
metadata := tx.Metadata()
metadata.Put(mempool.EstimateFeeDatabaseKey, s.feeEstimator.Save())
return nil
})
// Signal the remaining goroutines to quit.
close(s.quit)
2013-08-06 23:55:22 +02:00
return nil
}
// WaitForShutdown blocks until the main listener and peer handlers are stopped.
func (s *server) WaitForShutdown() {
s.wg.Wait()
}
// ScheduleShutdown schedules a server shutdown after the specified duration.
// It also dynamically adjusts how often to warn the server is going down based
// on remaining duration.
func (s *server) ScheduleShutdown(duration time.Duration) {
// Don't schedule shutdown more than once.
if atomic.AddInt32(&s.shutdownSched, 1) != 1 {
2013-08-06 23:55:22 +02:00
return
}
srvrLog.Warnf("Server shutdown in %v", duration)
2013-08-06 23:55:22 +02:00
go func() {
remaining := duration
tickDuration := dynamicTickDuration(remaining)
done := time.After(remaining)
ticker := time.NewTicker(tickDuration)
out:
for {
select {
case <-done:
ticker.Stop()
s.Stop()
break out
case <-ticker.C:
remaining = remaining - tickDuration
if remaining < time.Second {
continue
}
// Change tick duration dynamically based on remaining time.
newDuration := dynamicTickDuration(remaining)
if tickDuration != newDuration {
tickDuration = newDuration
ticker.Stop()
ticker = time.NewTicker(tickDuration)
}
srvrLog.Warnf("Server shutdown in %v", remaining)
2013-08-06 23:55:22 +02:00
}
}
}()
}
// parseListeners determines whether each listen address is IPv4 and IPv6 and
// returns a slice of appropriate net.Addrs to listen on with TCP. It also
// properly detects addresses which apply to "all interfaces" and adds the
// address as both IPv4 and IPv6.
func parseListeners(addrs []string) ([]net.Addr, error) {
netAddrs := make([]net.Addr, 0, len(addrs)*2)
for _, addr := range addrs {
host, _, err := net.SplitHostPort(addr)
if err != nil {
2013-11-14 02:43:36 +01:00
// Shouldn't happen due to already being normalized.
return nil, err
}
2013-11-14 02:43:36 +01:00
// Empty host or host of * on plan9 is both IPv4 and IPv6.
if host == "" || (host == "*" && runtime.GOOS == "plan9") {
netAddrs = append(netAddrs, simpleAddr{net: "tcp4", addr: addr})
netAddrs = append(netAddrs, simpleAddr{net: "tcp6", addr: addr})
continue
}
// Strip IPv6 zone id if present since net.ParseIP does not
// handle it.
zoneIndex := strings.LastIndex(host, "%")
if zoneIndex > 0 {
host = host[:zoneIndex]
}
// Parse the IP.
ip := net.ParseIP(host)
if ip == nil {
return nil, fmt.Errorf("'%s' is not a valid IP address", host)
}
2013-11-14 02:43:36 +01:00
// To4 returns nil when the IP is not an IPv4 address, so use
// this determine the address type.
if ip.To4() == nil {
netAddrs = append(netAddrs, simpleAddr{net: "tcp6", addr: addr})
} else {
netAddrs = append(netAddrs, simpleAddr{net: "tcp4", addr: addr})
}
}
return netAddrs, nil
}
func (s *server) upnpUpdateThread() {
// Go off immediately to prevent code duplication, thereafter we renew
// lease every 15 minutes.
timer := time.NewTimer(0 * time.Second)
lport, _ := strconv.ParseInt(activeNetParams.DefaultPort, 10, 16)
first := true
out:
for {
select {
case <-timer.C:
2016-11-16 00:43:43 +01:00
// TODO: pick external port more cleverly
// TODO: know which ports we are listening to on an external net.
// TODO: if specific listen port doesn't work then ask for wildcard
// listen port?
// XXX this assumes timeout is in seconds.
listenPort, err := s.nat.AddPortMapping("tcp", int(lport), int(lport),
"btcd listen port", 20*60)
if err != nil {
srvrLog.Warnf("can't add UPnP port mapping: %v", err)
}
if first && err == nil {
2016-11-16 00:43:43 +01:00
// TODO: look this up periodically to see if upnp domain changed
// and so did ip.
externalip, err := s.nat.GetExternalAddress()
if err != nil {
srvrLog.Warnf("UPnP can't get external address: %v", err)
continue out
}
na := wire.NewNetAddressIPPort(externalip, uint16(listenPort),
s.services)
err = s.addrManager.AddLocalAddress(na, addrmgr.UpnpPrio)
if err != nil {
// XXX DeletePortMapping?
}
srvrLog.Warnf("Successfully bound via UPnP to %s", addrmgr.NetAddressKey(na))
first = false
}
timer.Reset(time.Minute * 15)
case <-s.quit:
break out
}
}
timer.Stop()
if err := s.nat.DeletePortMapping("tcp", int(lport), int(lport)); err != nil {
srvrLog.Warnf("unable to remove UPnP port mapping: %v", err)
} else {
srvrLog.Debugf("successfully disestablished UPnP port mapping")
}
s.wg.Done()
}
// setupRPCListeners returns a slice of listeners that are configured for use
// with the RPC server depending on the configuration settings for listen
// addresses and TLS.
func setupRPCListeners() ([]net.Listener, error) {
// Setup TLS if not disabled.
listenFunc := net.Listen
if !cfg.DisableTLS {
// Generate the TLS cert and key file if both don't already
// exist.
if !fileExists(cfg.RPCKey) && !fileExists(cfg.RPCCert) {
err := genCertPair(cfg.RPCCert, cfg.RPCKey)
if err != nil {
return nil, err
}
}
keypair, err := tls.LoadX509KeyPair(cfg.RPCCert, cfg.RPCKey)
if err != nil {
return nil, err
}
tlsConfig := tls.Config{
Certificates: []tls.Certificate{keypair},
MinVersion: tls.VersionTLS12,
}
// Change the standard net.Listen function to the tls one.
listenFunc = func(net string, laddr string) (net.Listener, error) {
return tls.Listen(net, laddr, &tlsConfig)
}
}
netAddrs, err := parseListeners(cfg.RPCListeners)
if err != nil {
return nil, err
}
listeners := make([]net.Listener, 0, len(netAddrs))
for _, addr := range netAddrs {
listener, err := listenFunc(addr.Network(), addr.String())
if err != nil {
rpcsLog.Warnf("Can't listen on %s: %v", addr, err)
continue
}
listeners = append(listeners, listener)
}
return listeners, nil
}
2013-08-06 23:55:22 +02:00
// newServer returns a new btcd server configured to listen on addr for the
// bitcoin network type specified by chainParams. Use start to begin accepting
2013-08-06 23:55:22 +02:00
// connections from peers.
func newServer(listenAddrs, agentBlacklist, agentWhitelist []string,
db database.DB, chainParams *chaincfg.Params,
interrupt <-chan struct{}) (*server, error) {
startupTime := time.Now()
services := defaultServices
if cfg.NoPeerBloomFilters {
services &^= wire.SFNodeBloom
}
if cfg.NoCFilters {
services &^= wire.SFNodeCF
}
amgr := addrmgr.New(cfg.DataDir, btcdLookup)
var listeners []net.Listener
var nat NAT
if !cfg.DisableListen {
var err error
listeners, nat, err = initListeners(amgr, listenAddrs, services)
if err != nil {
return nil, err
}
if len(listeners) == 0 {
golint -min_confidence=0.3 . This commits removes a number of golint warnings. There is a class of warnings which I can't fix due to unsufficient knowledge of the domain at this point. These are listed here: addrmanager.go:907:1: comment on exported method AddrManager.Attempt should be of the form "Attempt ..." addrmanager.go:1048:1: exported function RFC1918 should have comment or be unexported addrmanager.go:1058:1: exported function RFC3849 should have comment or be unexported addrmanager.go:1065:1: exported function RFC3927 should have comment or be unexported addrmanager.go:1073:1: exported function RFC3964 should have comment or be unexported addrmanager.go:1081:1: exported function RFC4193 should have comment or be unexported addrmanager.go:1089:1: exported function RFC4380 should have comment or be unexported addrmanager.go:1097:1: exported function RFC4843 should have comment or be unexported addrmanager.go:1105:1: exported function RFC4862 should have comment or be unexported addrmanager.go:1113:1: exported function RFC6052 should have comment or be unexported addrmanager.go:1121:1: exported function RFC6145 should have comment or be unexported addrmanager.go:1128:1: exported function Tor should have comment or be unexported addrmanager.go:1143:1: exported function Local should have comment or be unexported addrmanager.go:1228:2: exported const InterfacePrio should have comment (or a comment on this block) or be unexported discovery.go:26:2: exported var ErrTorInvalidAddressResponse should have comment or be unexported limits/limits_unix.go:19:1: exported function SetLimits should have comment or be unexported limits/limits_windows.go:7:1: exported function SetLimits should have comment or be unexported util/dropafter/dropafter.go:22:6: exported type ShaHash should have comment or be unexported util/dropafter/dropafter.go:38:2: exported const ArgSha should have comment (or a comment on this block) or be unexported util/dropafter/dropafter.go:128:5: exported var ErrBadShaPrefix should have comment or be unexported util/dropafter/dropafter.go:129:5: exported var ErrBadShaLen should have comment or be unexported util/dropafter/dropafter.go:130:5: exported var ErrBadShaChar should have comment or be unexported util/showblock/showblock.go:24:6: exported type ShaHash should have comment or be unexported util/showblock/showblock.go:46:2: exported const ArgSha should have comment (or a comment on this block) or be unexported util/showblock/showblock.go:163:1: exported function DumpBlock should have comment or be unexported util/showblock/showblock.go:211:5: exported var ErrBadShaPrefix should have comment or be unexported util/showblock/showblock.go:212:5: exported var ErrBadShaLen should have comment or be unexported util/showblock/showblock.go:213:5: exported var ErrBadShaChar should have comment or be unexported
2014-07-02 16:25:42 +02:00
return nil, errors.New("no valid listen address")
}
2013-08-06 23:55:22 +02:00
}
if len(agentBlacklist) > 0 {
srvrLog.Infof("User-agent blacklist %s", agentBlacklist)
}
if len(agentWhitelist) > 0 {
srvrLog.Infof("User-agent whitelist %s", agentWhitelist)
}
2013-08-06 23:55:22 +02:00
s := server{
chainParams: chainParams,
addrManager: amgr,
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
newPeers: make(chan *serverPeer, cfg.MaxPeers),
donePeers: make(chan *serverPeer, cfg.MaxPeers),
banPeers: make(chan *serverPeer, cfg.MaxPeers),
query: make(chan interface{}),
relayInv: make(chan relayMsg, cfg.MaxPeers),
broadcast: make(chan broadcastMsg, cfg.MaxPeers),
quit: make(chan struct{}),
modifyRebroadcastInv: make(chan interface{}),
peerHeightsUpdate: make(chan updatePeerHeightsMsg),
nat: nat,
db: db,
timeSource: blockchain.NewMedianTime(),
services: services,
Integrate a valid ECDSA signature cache into btcd Introduce an ECDSA signature verification into btcd in order to mitigate a certain DoS attack and as a performance optimization. The benefits of SigCache are two fold. Firstly, usage of SigCache mitigates a DoS attack wherein an attacker causes a victim's client to hang due to worst-case behavior triggered while processing attacker crafted invalid transactions. A detailed description of the mitigated DoS attack can be found here: https://bitslog.wordpress.com/2013/01/23/fixed-bitcoin-vulnerability-explanation-why-the-signature-cache-is-a-dos-protection/ Secondly, usage of the SigCache introduces a signature verification optimization which speeds up the validation of transactions within a block, if they've already been seen and verified within the mempool. The server itself manages the sigCache instance. The blockManager and txMempool respectively now receive pointers to the created sigCache instance. All read (sig triplet existence) operations on the sigCache will not block unless a separate goroutine is adding an entry (writing) to the sigCache. GetBlockTemplate generation now also utilizes the sigCache in order to avoid unnecessarily double checking signatures when generating a template after previously accepting a txn to the mempool. Consequently, the CPU miner now also employs the same optimization. The maximum number of entries for the sigCache has been introduced as a config parameter in order to allow users to configure the amount of memory consumed by this new additional caching.
2015-09-25 01:22:00 +02:00
sigCache: txscript.NewSigCache(cfg.SigCacheMaxSize),
hashCache: txscript.NewHashCache(cfg.SigCacheMaxSize),
cfCheckptCaches: make(map[wire.FilterType][]cfHeaderKV),
agentBlacklist: agentBlacklist,
agentWhitelist: agentWhitelist,
2013-08-06 23:55:22 +02:00
}
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
// Create the transaction and address indexes if needed.
//
// CAUTION: the txindex needs to be first in the indexes array because
// the addrindex uses data from the txindex during catchup. If the
// addrindex is run first, it may not have the transactions from the
// current block indexed.
var indexes []indexers.Indexer
if cfg.TxIndex || cfg.AddrIndex {
// Enable transaction index if address index is enabled since it
// requires it.
if !cfg.TxIndex {
indxLog.Infof("Transaction index enabled because it " +
"is required by the address index")
cfg.TxIndex = true
} else {
indxLog.Info("Transaction index is enabled")
}
s.txIndex = indexers.NewTxIndex(db)
indexes = append(indexes, s.txIndex)
}
if cfg.AddrIndex {
indxLog.Info("Address index is enabled")
s.addrIndex = indexers.NewAddrIndex(db, chainParams)
indexes = append(indexes, s.addrIndex)
}
if !cfg.NoCFilters {
indxLog.Info("Committed filter index is enabled")
s.cfIndex = indexers.NewCfIndex(db, chainParams)
indexes = append(indexes, s.cfIndex)
2017-01-12 12:20:16 +01:00
}
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
// Create an index manager if any of the optional indexes are enabled.
var indexManager blockchain.IndexManager
if len(indexes) > 0 {
indexManager = indexers.NewManager(db, indexes)
}
// Merge given checkpoints with the default ones unless they are disabled.
var checkpoints []chaincfg.Checkpoint
if !cfg.DisableCheckpoints {
checkpoints = mergeCheckpoints(s.chainParams.Checkpoints, cfg.addCheckpoints)
}
// Create a new block chain instance with the appropriate configuration.
var err error
s.chain, err = blockchain.New(&blockchain.Config{
DB: s.db,
Interrupt: interrupt,
ChainParams: s.chainParams,
Checkpoints: checkpoints,
TimeSource: s.timeSource,
SigCache: s.sigCache,
IndexManager: indexManager,
HashCache: s.hashCache,
})
2013-08-06 23:55:22 +02:00
if err != nil {
return nil, err
}
// Search for a FeeEstimator state in the database. If none can be found
// or if it cannot be loaded, create a new one.
db.Update(func(tx database.Tx) error {
metadata := tx.Metadata()
feeEstimationData := metadata.Get(mempool.EstimateFeeDatabaseKey)
if feeEstimationData != nil {
// delete it from the database so that we don't try to restore the
// same thing again somehow.
metadata.Delete(mempool.EstimateFeeDatabaseKey)
// If there is an error, log it and make a new fee estimator.
var err error
s.feeEstimator, err = mempool.RestoreFeeEstimator(feeEstimationData)
if err != nil {
peerLog.Errorf("Failed to restore fee estimator %v", err)
}
}
return nil
})
// If no feeEstimator has been found, or if the one that has been found
// is behind somehow, create a new one and start over.
if s.feeEstimator == nil || s.feeEstimator.LastKnownHeight() != s.chain.BestSnapshot().Height {
s.feeEstimator = mempool.NewFeeEstimator(
mempool.DefaultEstimateFeeMaxRollback,
mempool.DefaultEstimateFeeMinRegisteredBlocks)
}
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
txC := mempool.Config{
Policy: mempool.Policy{
DisableRelayPriority: cfg.NoRelayPriority,
AcceptNonStd: cfg.RelayNonStd,
FreeTxRelayLimit: cfg.FreeTxRelayLimit,
MaxOrphanTxs: cfg.MaxOrphanTxs,
MaxOrphanTxSize: defaultMaxOrphanTxSize,
MaxSigOpCostPerTx: blockchain.MaxBlockSigOpsCost / 4,
MinRelayTxFee: cfg.minRelayTxFee,
MaxTxVersion: 2,
RejectReplacement: cfg.RejectReplacement,
},
ChainParams: chainParams,
FetchUtxoView: s.chain.FetchUtxoView,
BestHeight: func() int32 { return s.chain.BestSnapshot().Height },
MedianTimePast: func() time.Time { return s.chain.BestSnapshot().MedianTime },
CalcSequenceLock: func(tx *btcutil.Tx, view *blockchain.UtxoViewpoint) (*blockchain.SequenceLock, error) {
return s.chain.CalcSequenceLock(tx, view, true)
},
IsDeploymentActive: s.chain.IsDeploymentActive,
SigCache: s.sigCache,
HashCache: s.hashCache,
AddrIndex: s.addrIndex,
FeeEstimator: s.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
s.txMemPool = mempool.New(&txC)
s.syncManager, err = netsync.New(&netsync.Config{
PeerNotifier: &s,
Chain: s.chain,
TxMemPool: s.txMemPool,
ChainParams: s.chainParams,
DisableCheckpoints: cfg.DisableCheckpoints,
MaxPeers: cfg.MaxPeers,
FeeEstimator: s.feeEstimator,
})
if err != nil {
return nil, err
}
// Create the mining policy and block template generator based on the
// configuration options.
//
// NOTE: The CPU miner relies on the mempool, so the mempool has to be
// created before calling the function to create the CPU miner.
policy := mining.Policy{
BlockMinWeight: cfg.BlockMinWeight,
BlockMaxWeight: cfg.BlockMaxWeight,
BlockMinSize: cfg.BlockMinSize,
BlockMaxSize: cfg.BlockMaxSize,
BlockPrioritySize: cfg.BlockPrioritySize,
TxMinFreeFee: cfg.minRelayTxFee,
}
blockTemplateGenerator := mining.NewBlkTmplGenerator(&policy,
s.chainParams, s.txMemPool, s.chain, s.timeSource,
s.sigCache, s.hashCache)
s.cpuMiner = cpuminer.New(&cpuminer.Config{
ChainParams: chainParams,
BlockTemplateGenerator: blockTemplateGenerator,
MiningAddrs: cfg.miningAddrs,
ProcessBlock: s.syncManager.ProcessBlock,
ConnectedCount: s.ConnectedCount,
IsCurrent: s.syncManager.IsCurrent,
})
// Only setup a function to return new addresses to connect to when
// not running in connect-only mode. The simulation network is always
// in connect-only mode since it is only intended to connect to
// specified peers and actively avoid advertising and connecting to
// discovered peers in order to prevent it from becoming a public test
// network.
var newAddressFunc func() (net.Addr, error)
if !cfg.SimNet && len(cfg.ConnectPeers) == 0 {
newAddressFunc = func() (net.Addr, error) {
for tries := 0; tries < 100; tries++ {
addr := s.addrManager.GetAddress()
if addr == nil {
break
}
// Address will not be invalid, local or unroutable
// because addrmanager rejects those on addition.
// Just check that we don't already have an address
// in the same group so that we are not connecting
// to the same network segment at the expense of
// others.
key := addrmgr.GroupKey(addr.NetAddress())
if s.OutboundGroupCount(key) != 0 {
continue
}
// only allow recent nodes (10mins) after we failed 30
// times
if tries < 30 && time.Since(addr.LastAttempt()) < 10*time.Minute {
continue
}
// allow nondefault ports after 50 failed tries.
if tries < 50 && fmt.Sprintf("%d", addr.NetAddress().Port) !=
activeNetParams.DefaultPort {
continue
}
// Mark an attempt for the valid address.
s.addrManager.Attempt(addr.NetAddress())
addrString := addrmgr.NetAddressKey(addr.NetAddress())
return addrStringToNetAddr(addrString)
}
return nil, errors.New("no valid connect address")
}
}
// Create a connection manager.
targetOutbound := defaultTargetOutbound
if cfg.MaxPeers < targetOutbound {
targetOutbound = cfg.MaxPeers
}
cmgr, err := connmgr.New(&connmgr.Config{
Listeners: listeners,
OnAccept: s.inboundPeerConnected,
RetryDuration: connectionRetryInterval,
TargetOutbound: uint32(targetOutbound),
Dial: btcdDial,
OnConnection: s.outboundPeerConnected,
GetNewAddress: newAddressFunc,
})
if err != nil {
return nil, err
}
s.connManager = cmgr
// Start up persistent peers.
permanentPeers := cfg.ConnectPeers
if len(permanentPeers) == 0 {
permanentPeers = cfg.AddPeers
}
for _, addr := range permanentPeers {
netAddr, err := addrStringToNetAddr(addr)
if err != nil {
return nil, err
}
go s.connManager.Connect(&connmgr.ConnReq{
Addr: netAddr,
Permanent: true,
})
}
if !cfg.DisableRPC {
// Setup listeners for the configured RPC listen addresses and
// TLS settings.
rpcListeners, err := setupRPCListeners()
if err != nil {
return nil, err
}
if len(rpcListeners) == 0 {
return nil, errors.New("RPCS: No valid listen address")
}
s.rpcServer, err = newRPCServer(&rpcserverConfig{
2017-11-13 23:39:16 +01:00
Listeners: rpcListeners,
StartupTime: startupTime.Unix(),
2017-11-13 23:39:16 +01:00
ConnMgr: &rpcConnManager{&s},
SyncMgr: &rpcSyncMgr{&s, s.syncManager},
TimeSource: s.timeSource,
Chain: s.chain,
ChainParams: chainParams,
DB: db,
TxMemPool: s.txMemPool,
Generator: blockTemplateGenerator,
CPUMiner: s.cpuMiner,
TxIndex: s.txIndex,
AddrIndex: s.addrIndex,
CfIndex: s.cfIndex,
FeeEstimator: s.feeEstimator,
})
2013-08-06 23:55:22 +02:00
if err != nil {
return nil, err
}
// Signal process shutdown when the RPC server requests it.
go func() {
<-s.rpcServer.RequestedProcessShutdown()
shutdownRequestChannel <- struct{}{}
}()
2013-08-06 23:55:22 +02:00
}
Implement a built-in concurrent CPU miner. This commit implements a built-in concurrent CPU miner that can be enabled with the combination of the --generate and --miningaddr options. The --blockminsize, --blockmaxsize, and --blockprioritysize configuration options wich already existed prior to this commit control the block template generation and hence affect blocks mined via the new CPU miner. The following is a quick overview of the changes and design: - Starting btcd with --generate and no addresses specified via --miningaddr will give an error and exit immediately - Makes use of multiple worker goroutines which independently create block templates, solve them, and submit the solved blocks - The default number of worker threads are based on the number of processor cores in the system and can be dynamically changed at run-time - There is a separate speed monitor goroutine used to collate periodic updates from the workers to calculate overall hashing speed - The current mining state, number of workers, and hashes per second can be queried - Updated sample-btcd.conf file has been updated to include the coin generation (mining) settings - Updated doc.go for the new command line options In addition the old --getworkkey option is now deprecated in favor of the new --miningaddr option. This was changed for a few reasons: - There is no reason to have a separate list of keys for getwork and CPU mining - getwork is deprecated and will be going away in the future so that means the --getworkkey flag will also be going away - Having the work 'key' in the option can be confused with wanting a private key while --miningaddr make it a little more clear it is an address that is required Closes #137. Reviewed by @jrick.
2014-06-12 03:09:38 +02:00
2013-08-06 23:55:22 +02:00
return &s, nil
}
// initListeners initializes the configured net listeners and adds any bound
// addresses to the address manager. Returns the listeners and a NAT interface,
// which is non-nil if UPnP is in use.
func initListeners(amgr *addrmgr.AddrManager, listenAddrs []string, services wire.ServiceFlag) ([]net.Listener, NAT, error) {
// Listen for TCP connections at the configured addresses
netAddrs, err := parseListeners(listenAddrs)
if err != nil {
return nil, nil, err
}
listeners := make([]net.Listener, 0, len(netAddrs))
for _, addr := range netAddrs {
listener, err := net.Listen(addr.Network(), addr.String())
if err != nil {
srvrLog.Warnf("Can't listen on %s: %v", addr, err)
continue
}
listeners = append(listeners, listener)
}
var nat NAT
if len(cfg.ExternalIPs) != 0 {
defaultPort, err := strconv.ParseUint(activeNetParams.DefaultPort, 10, 16)
if err != nil {
srvrLog.Errorf("Can not parse default port %s for active chain: %v",
activeNetParams.DefaultPort, err)
return nil, nil, err
}
for _, sip := range cfg.ExternalIPs {
eport := uint16(defaultPort)
host, portstr, err := net.SplitHostPort(sip)
if err != nil {
// no port, use default.
host = sip
} else {
port, err := strconv.ParseUint(portstr, 10, 16)
if err != nil {
srvrLog.Warnf("Can not parse port from %s for "+
"externalip: %v", sip, err)
continue
}
eport = uint16(port)
}
na, err := amgr.HostToNetAddress(host, eport, services)
if err != nil {
srvrLog.Warnf("Not adding %s as externalip: %v", sip, err)
continue
}
err = amgr.AddLocalAddress(na, addrmgr.ManualPrio)
if err != nil {
amgrLog.Warnf("Skipping specified external IP: %v", err)
}
}
} else {
if cfg.Upnp {
var err error
nat, err = Discover()
if err != nil {
srvrLog.Warnf("Can't discover upnp: %v", err)
}
// nil nat here is fine, just means no upnp on network.
}
// Add bound addresses to address manager to be advertised to peers.
for _, listener := range listeners {
addr := listener.Addr().String()
err := addLocalAddress(amgr, addr, services)
if err != nil {
amgrLog.Warnf("Skipping bound address %s: %v", addr, err)
}
}
}
return listeners, nat, nil
}
// addrStringToNetAddr takes an address in the form of 'host:port' and returns
// a net.Addr which maps to the original address with any host names resolved
// to IP addresses. It also handles tor addresses properly by returning a
// net.Addr that encapsulates the address.
func addrStringToNetAddr(addr string) (net.Addr, error) {
host, strPort, err := net.SplitHostPort(addr)
if err != nil {
return nil, err
}
port, err := strconv.Atoi(strPort)
if err != nil {
return nil, err
}
// Skip if host is already an IP address.
if ip := net.ParseIP(host); ip != nil {
return &net.TCPAddr{
IP: ip,
Port: port,
}, nil
}
// Tor addresses cannot be resolved to an IP, so just return an onion
// address instead.
if strings.HasSuffix(host, ".onion") {
if cfg.NoOnion {
return nil, errors.New("tor has been disabled")
}
return &onionAddr{addr: addr}, nil
}
// Attempt to look up an IP address associated with the parsed host.
ips, err := btcdLookup(host)
if err != nil {
return nil, err
}
if len(ips) == 0 {
return nil, fmt.Errorf("no addresses found for %s", host)
}
return &net.TCPAddr{
IP: ips[0],
Port: port,
}, nil
}
// addLocalAddress adds an address that this node is listening on to the
// address manager so that it may be relayed to peers.
func addLocalAddress(addrMgr *addrmgr.AddrManager, addr string, services wire.ServiceFlag) error {
host, portStr, err := net.SplitHostPort(addr)
if err != nil {
return err
}
port, err := strconv.ParseUint(portStr, 10, 16)
if err != nil {
return err
}
if ip := net.ParseIP(host); ip != nil && ip.IsUnspecified() {
// If bound to unspecified address, advertise all local interfaces
addrs, err := net.InterfaceAddrs()
if err != nil {
return err
}
for _, addr := range addrs {
ifaceIP, _, err := net.ParseCIDR(addr.String())
if err != nil {
continue
}
// If bound to 0.0.0.0, do not add IPv6 interfaces and if bound to
// ::, do not add IPv4 interfaces.
if (ip.To4() == nil) != (ifaceIP.To4() == nil) {
continue
}
netAddr := wire.NewNetAddressIPPort(ifaceIP, uint16(port), services)
addrMgr.AddLocalAddress(netAddr, addrmgr.BoundPrio)
}
} else {
netAddr, err := addrMgr.HostToNetAddress(host, uint16(port), services)
if err != nil {
return err
}
addrMgr.AddLocalAddress(netAddr, addrmgr.BoundPrio)
}
return nil
}
2013-08-06 23:55:22 +02:00
// dynamicTickDuration is a convenience function used to dynamically choose a
// tick duration based on remaining time. It is primarily used during
// server shutdown to make shutdown warnings more frequent as the shutdown time
// approaches.
func dynamicTickDuration(remaining time.Duration) time.Duration {
switch {
case remaining <= time.Second*5:
return time.Second
case remaining <= time.Second*15:
return time.Second * 5
case remaining <= time.Minute:
return time.Second * 15
case remaining <= time.Minute*5:
return time.Minute
case remaining <= time.Minute*15:
return time.Minute * 5
case remaining <= time.Hour:
return time.Minute * 15
}
return time.Hour
}
2017-08-31 15:59:43 +02:00
// isWhitelisted returns whether the IP address is included in the whitelisted
// networks and IPs.
func isWhitelisted(addr net.Addr) bool {
if len(cfg.whitelists) == 0 {
return false
}
host, _, err := net.SplitHostPort(addr.String())
if err != nil {
srvrLog.Warnf("Unable to SplitHostPort on '%s': %v", addr, err)
return false
}
ip := net.ParseIP(host)
if ip == nil {
srvrLog.Warnf("Unable to parse IP '%s'", addr)
return false
}
for _, ipnet := range cfg.whitelists {
if ipnet.Contains(ip) {
return true
}
}
return false
}
// checkpointSorter implements sort.Interface to allow a slice of checkpoints to
// be sorted.
type checkpointSorter []chaincfg.Checkpoint
// Len returns the number of checkpoints in the slice. It is part of the
// sort.Interface implementation.
func (s checkpointSorter) Len() int {
return len(s)
}
// Swap swaps the checkpoints at the passed indices. It is part of the
// sort.Interface implementation.
func (s checkpointSorter) Swap(i, j int) {
s[i], s[j] = s[j], s[i]
}
// Less returns whether the checkpoint with index i should sort before the
// checkpoint with index j. It is part of the sort.Interface implementation.
func (s checkpointSorter) Less(i, j int) bool {
return s[i].Height < s[j].Height
}
// mergeCheckpoints returns two slices of checkpoints merged into one slice
// such that the checkpoints are sorted by height. In the case the additional
// checkpoints contain a checkpoint with the same height as a checkpoint in the
// default checkpoints, the additional checkpoint will take precedence and
// overwrite the default one.
func mergeCheckpoints(defaultCheckpoints, additional []chaincfg.Checkpoint) []chaincfg.Checkpoint {
// Create a map of the additional checkpoints to remove duplicates while
// leaving the most recently-specified checkpoint.
extra := make(map[int32]chaincfg.Checkpoint)
for _, checkpoint := range additional {
extra[checkpoint.Height] = checkpoint
}
// Add all default checkpoints that do not have an override in the
// additional checkpoints.
numDefault := len(defaultCheckpoints)
checkpoints := make([]chaincfg.Checkpoint, 0, numDefault+len(extra))
for _, checkpoint := range defaultCheckpoints {
if _, exists := extra[checkpoint.Height]; !exists {
checkpoints = append(checkpoints, checkpoint)
}
}
// Append the additional checkpoints and return the sorted results.
for _, checkpoint := range extra {
checkpoints = append(checkpoints, checkpoint)
}
sort.Sort(checkpointSorter(checkpoints))
return checkpoints
}
// HasUndesiredUserAgent determines whether the server should continue to pursue
// a connection with this peer based on its advertised user agent. It performs
// the following steps:
// 1) Reject the peer if it contains a blacklisted agent.
// 2) If no whitelist is provided, accept all user agents.
// 3) Accept the peer if it contains a whitelisted agent.
// 4) Reject all other peers.
func (sp *serverPeer) HasUndesiredUserAgent(blacklistedAgents,
whitelistedAgents []string) bool {
agent := sp.UserAgent()
// First, if peer's user agent contains any blacklisted substring, we
// will ignore the connection request.
for _, blacklistedAgent := range blacklistedAgents {
if strings.Contains(agent, blacklistedAgent) {
srvrLog.Debugf("Ignoring peer %s, user agent "+
"contains blacklisted user agent: %s", sp,
agent)
return true
}
}
// If no whitelist is provided, we will accept all user agents.
if len(whitelistedAgents) == 0 {
return false
}
// Peer's user agent passed blacklist. Now check to see if it contains
// one of our whitelisted user agents, if so accept.
for _, whitelistedAgent := range whitelistedAgents {
if strings.Contains(agent, whitelistedAgent) {
return false
}
}
// Otherwise, the peer's user agent was not included in our whitelist.
// Ignore just in case it could stall the initial block download.
srvrLog.Debugf("Ignoring peer %s, user agent: %s not found in "+
"whitelist", sp, agent)
return true
}