* Remove un-needed #includes (what is the policy?)
Data was duplicated in masternode_info_t and CMasternode classes:
* CMasternode is changed to inherit from masternode_info_t
so the data members are inherited rather than repeated
(also inherits unrepeated nTimeLastPing and fInfoValid members;
this slight intrusiveness made up for in simplicity).
* Use in-class member initializers (C++11) for defaults,
so only non-default initializers are required in the lists.
Allows to shorten repetitous constructor initializer lists.
This makes checking for uninitialized data simpler.
* Default constructors are defined as "= default;" if possible.
* masternode_info_t is changed to behave like an aggregate
(but requires over-complicated constructors until c++14).
There are pros and cons here - aggregate initialization
is convenient but implicit).
* Removed user-defined swap functions.
They appear to only be used in operator= definitions,
using the copy-in,swap-and-return idiom:
* Default operator=, where possible.
* Move in class `friend bool operator==` out-of-class.
* Change sync process:
- IsBlockchainSynced(): drop CheckNodeHeight() and all complicated code, use fInitialDownload in UpdatedBlockTip() to switch initial states
- ProcessTick(): detect sleep mode like it was in IsBlockchainSynced(), not by number of masternodes
* Changes for sync in governance:
- do not keep sync alive on ConfirmInventoryRequest()
- skip some governance actions until we are synced to some level
* do not run CMasternodeMan::UpdateLastPaid() until winners list is synced
* start syncing mn list on the same node right after requesting sporks
* replace nTimeLast<Asset> with the unified nTimeLastBumped, bump on UpdatedBlockTip
* fix comments and LogPrintf-s
* remove excessive MASTERNODE_SYNC_IBD
* a bit more descriptive BumpAssetLastTime in few cases
* net: a few small cleanups before replacing boost threads
- Drop the interruption point directly after the pnode allocation. This would
be leaky if hit.
- Rearrange thread creation so that the socket handler comes first
* net: add CThreadInterrupt and InterruptibleSleep
* net: make net interruptible
Also now that net threads are interruptible, switch them to use std
threads/binds/mutexes/condvars.
* net: make net processing interruptible
* net: remove thread_interrupted catch
This is now a std::thread, so there's no hope of catching a boost interruption
point.
* net: make proxy receives interruptible
* net: misc header cleanups
* Remove orphan state wipe from UnloadBlockIndex.
As orphan state is now "network state", like in
d6ea737be1,
UnloadBlockIndex is only used during init if we end up reindexing
to clear our block state so that we can start over. However, at
that time no connections have been brought up as CConnman hasn't
been started yet, so all of the network processing state logic is
empty when its called.
* Move network-msg-processing code out of main to its own file
* Rename the remaining main.{h,cpp} to validation.{h,cpp}
* net: Consistent checksum handling
In principle, the checksums of P2P packets are simply 4-byte blobs which
are the first four bytes of SHA256(SHA256(payload)).
Currently they are handled as little-endian 32-bit integers half of the
time, as blobs the other half, sometimes copying the one to the other,
resulting in somewhat confused code.
This PR changes the handling to be consistent both at packet creation
and receiving, making it (I think) easier to understand.
* net: Hardcode protocol sizes and use fixed-size types
The P2P network uses a fixed protocol, these sizes shouldn't change
based on what happens to be the architecture.
* Expose AcceptBlockHeader through main.h
* Split ::HEADERS processing into two separate cs_main locks
This will allow NotifyHeaderTip to be called from an
AcceptBlockHeader wrapper function without holding cs_main.
* Use exposed ProcessNewBlockHeaders from ProcessMessages
* Remove pfrom parameter from ProcessNewBlock
This further decouples ProcessNewBlock from networking/peer logic.
* Replace CValidationState param in ProcessNewBlock with BlockChecked
* Move MarkBlockAsReceived out of ProcessNewMessage
* Remove network state wipe from UnloadBlockIndex.
UnloadBlockIndex is only used during init if we end up reindexing
to clear our block state so that we can start over. However, at
that time no connections have been brought up as CConnman hasn't
been started yet, so all of the network processing state logic is
empty when its called.
Additionally, the initialization of the recentRejects set is moved
to InitPeerLogic.
* Move all calls to CheckBlockIndex out of net-processing logic
This will result in many more calls to CheckBlockIndex when
connecting a list of headers (eg in ::HEADERS messages processing)
but its only enabled in debug mode, and that should mostly just be
during IBD, so it should be OK.
* Move FlushStateToDisk call out of ProcessMessages::TX into ATMP
* Move nTimeBestReceived updating into net processing code
* Make validationinterface.UpdatedBlockTip more verbose
In anticipation of making all the callbacks out of block processing
flow through it. Note that vHashes will always have something in it
since pindexFork != pindexNewTip.
* Remove duplicate nBlocksEstimate cmp (we already checked IsIBD())
* Remove CConnman parameter from ProcessNewBlock/ActivateBestChain
* Remove SyncWithWallets wrapper function
* Move net-processing logic definitions together in main.h
* Use CValidationInterface from chain logic to notify peer logic
This adds a new CValidationInterface subclass, defined in main.h,
to receive notifications of UpdatedBlockTip and use that to push
blocks to peers, instead of doing it directly from
ActivateBestChain.
* Always call UpdatedBlockTip, even if blocks were only disconnected
* Use BlockChecked signal to send reject messages from mapBlockSource
In the case of (for example) an already-running bitcoind, the shutdown sequence
begins before CConnman has been created, leading to a null-pointer dereference
when g_connman->Stop() is called.
Instead, Just let the CConnman dtor take care of stopping.
* serialization: teach serializers variadics
Also add a variadic CDataStream ctor for ease-of-use.
* connman is in charge of pushing messages
The changes here are dense and subtle, but hopefully all is more explicit
than before.
- CConnman is now in charge of sending data rather than the nodes themselves.
This is necessary because many decisions need to be made with all nodes in
mind, and a model that requires the nodes calling up to their manager quickly
turns to spaghetti.
- The per-node-serializer (ssSend) has been replaced with a (quasi-)const
send-version. Since the send version for serialization can only change once
per connection, we now explicitly tag messages with INIT_PROTO_VERSION if
they are sent before the handshake. With this done, there's no need to lock
for access to nSendVersion.
Also, a new stream is used for each message, so there's no need to lock
during the serialization process.
- This takes care of accounting for optimistic sends, so the
nOptimisticBytesWritten hack can be removed.
- -dropmessagestest and -fuzzmessagestest have not been preserved, as I suspect
they haven't been used in years.
* net: switch all callers to connman for pushing messages
Drop all of the old stuff.
* drop the optimistic write counter hack
This is now handled properly in realtime.
* net: remove now-unused ssSend and Fuzz
* net: construct CNodeStates in place
* net: handle version push in InitializeNode
* net: Add fRelayTxes flag
Add a fRelayTxes to keep track of the relay transaction flag
we send to other peers.
* rpc: Add `relaytxes` flag to `getnetworkinfo`
Re-work of PR #7841 by dragongem45.
Closes#7771.
* net: move CBanDB and CAddrDB out of net.h/cpp
This will eventually solve a circular dependency
* net: Create CConnman to encapsulate p2p connections
* net: Move socket binding into CConnman
* net: move OpenNetworkConnection into CConnman
* net: move ban and addrman functions into CConnman
* net: Add oneshot functions to CConnman
* net: move added node functions to CConnman
* net: Add most functions needed for vNodes to CConnman
* net: handle nodesignals in CConnman
* net: Pass CConnection to wallet rather than using the global
* net: Add rpc error for missing/disabled p2p functionality
* net: Pass CConnman around as needed
* gui: add NodeID to the peer table
* net: create generic functor accessors and move vNodes to CConnman
* net: move whitelist functions into CConnman
* net: move nLastNodeId to CConnman
* net: move nLocalHostNonce to CConnman
This behavior seems to have been quite racy and broken.
Move nLocalHostNonce into CNode, and check received nonces against all
non-fully-connected nodes. If there's a match, assume we've connected
to ourself.
* net: move messageHandlerCondition to CConnman
* net: move send/recv statistics to CConnman
* net: move SendBufferSize/ReceiveFloodSize to CConnman
* net: move nLocalServices/nRelevantServices to CConnman
These are in-turn passed to CNode at connection time. This allows us to offer
different services to different peers (or test the effects of doing so).
* net: move semOutbound and semMasternodeOutbound to CConnman
* net: SocketSendData returns written size
* net: move max/max-outbound to CConnman
* net: Pass best block known height into CConnman
CConnman then passes the current best height into CNode at creation time.
This way CConnman/CNode have no dependency on main for height, and the signals
only move in one direction.
This also helps to prevent identity leakage a tiny bit. Before this change, an
attacker could theoretically make 2 connections on different interfaces. They
would connect fully on one, and only establish the initial connection on the
other. Once they receive a new block, they would relay it to your first
connection, and immediately commence the version handshake on the second. Since
the new block height is reflected immediately, they could attempt to learn
whether the two connections were correlated.
This is, of course, incredibly unlikely to work due to the small timings
involved and receipt from other senders. But it doesn't hurt to lock-in
nBestHeight at the time of connection, rather than letting the remote choose
the time.
* net: pass CClientUIInterface into CConnman
* net: Drop StartNode/StopNode and use CConnman directly
* net: Introduce CConnection::Options to avoid passing so many params
* net: add nSendBufferMaxSize/nReceiveFloodSize to CConnection::Options
* net: move vNodesDisconnected into CConnman
* Made the ForEachNode* functions in src/net.cpp more pragmatic and self documenting
* Convert ForEachNode* functions to take a templated function argument rather than a std::function to eliminate std::function overhead
* net: move MAX_FEELER_CONNECTIONS into connman
* lock cs_main for chainActive
ActivateBestChain uses chainActive after releasing the lock; reorder operations
to move all access to synchronized object into existing LOCK(cs_main) block.
* lock cs_main for State/Misbehaving
ProcessMessage calls State(...) and Misbehaving(...) without holding the
required lock; add LOCK(cs_main) blocks.
Tests if addresses are online or offline by briefly connecting to them. These short lived connections are referred to as feeler connections. Feeler connections are designed to increase the number of fresh online addresses in tried by selecting and connecting to addresses in new. One feeler connection is attempted on average once every two minutes.
This change was suggested as Countermeasure 4 in
Eclipse Attacks on Bitcoin’s Peer-to-Peer Network, Ethan Heilman,
Alison Kendler, Aviv Zohar, Sharon Goldberg. ePrint Archive Report
2015/263. March 2015.
As per meeting 2016-03-31
https://bitcoincore.org/en/meetings/2016/03/31/#bad-chain-alerts
The partition checker was producing huge number of false-positives
and was disabled in 0.12.1 on the understanding it would either be
fixed in 0.13 or removed entirely from master if not.
*** Dash specific note: ***
This check was disabled in Dash already.
* Rework addnode behaviour
* Use CNode::addeName to track whether a connection to a name is already open
* A new connection to a previously-connected by-name addednode is only opened when
the previous one closes (even if the name starts resolving to something else)
* At most one connection is opened per addednode (even if the name resolves to multiple)
* Unify the code between ThreadOpenAddedNodeConnections and getaddednodeinfo
* Information about open connections is always returned, and the dns argument becomes a dummy
* An IP address and inbound/outbound is only reported for the (at most 1) open connection
* Prevent duplicate connections where one is by name and another by ip
* Randomize name lookup result in ConnectSocketByName
* fix vulnerability with mapMasternodeOrphanObjects
The vulnerability is that a malicious node can send a lot of NetMsgType::MNGOVERNANCEOBJECT messages which refer to many arbitrary MN's. In this case, mapMasternodeOrphanObjects will grow unrestrictedly.
* MN collateral moved to governance-object.cpp; ban score applied to misbehaving nodes
* recursive locks removed
* check for the mn collateral code segregated to a separate function
* CheckCollateral implementation moved to cpp
This reverts commit 1f828f45ec.
The commit being reverted changed FindNode(const CService& addr)
to make no difference between nodes with the same IP address,
but different ports, but only for regtest network.
As functional tests run several nodes on different ports or the same
IP address (127.0.0.1), this eventually started breaking functional tests.
The only use for regtest network I know is for functional tests,
so it's time to revert that commit.
* fix issues with mapSeenGovernanceObjects
Removed seen-governance-objects optimization except for deleted objects. Otherwise some nodes can permanently lost proposals if they received them too early.
Beside of that there is a vulnerability with seen-governance-objects mechanism if malicious node send us a lot of invalid governance objects.
* mapSeenGovernanceObjects renamed to mapErasedGovernanceObjects
* current fixes
* use int64_t for expiration timestamp
* Add recently accepted blocks and txn to AttemptToEvictConnection.
This protects any not-already-protected peers who were the most
recent four to relay transactions and most recent four to send
blocks to us.
* Allow disconnecting a netgroup with only one member in eviction.
With the latest additions there are enough protective measures that
we can take the training wheels off.
* net: don't import std namespace
This file is about to be broken up into chunks and moved around. Drop the
namespace now rather than requiring other files to use it.
* net: remove unused set
* net: use the exposed GetNodeSignals() rather than g_signals directly
* net: make Ban/Unban/ClearBan functionality consistent
- Ban/Unban/ClearBan call uiInterface.BannedListChanged() as necessary
- Ban/Unban/ClearBan sync to disk if the operation is user-invoked
- Mark node for disconnection automatically when banning
- Lock cs_vNodes while setting disconnected
- Don't spin in a tight loop while setting disconnected
* net: No need to export DumpBanlist
DumpBanList currently does this:
- with lock: take a copy of the banmap
- perform I/O (write out the banmap)
- with lock: mark the banmap non-dirty
If a new ban is added during the I/O operation, it may never be persisted to
disk.
Reorder operations so that the data to be persisted cannot be older than the
time at which the banmap was marked non-dirty.
* Make ProcessNewBlock dbp const and update comment
* Switch reindexing to AcceptBlock in-loop and ActivateBestChain afterwards
* Optimize ActivateBestChain for long chains
* Add -reindex-chainstate that does not rebuild block index
* Report reindexing progress in GUI
