1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250
//! Copyright (c) 2023 MASSA LABS <info@massa.net>
//! This file deals with the announcement of block headers to other nodes
//! in order to propagate the blocks from our node to other nodes.
//! It also manages peer banning for invalid blocks detected by consensus.
//!
//! The block propagation system works in the following way:
//! * a node announces the headers of blocks to its neighbor nodes
//! * the neighbor nodes that need that block then ask our Retrieval process for it
//!
//! Here we need to announce block headers to other nodes that haven't sene them,
//! and keep the blocks alive long enough for our peers to be able to retrieve them from us.
use super::{
cache::SharedBlockCache, commands_propagation::BlockHandlerPropagationCommand,
BlockMessageSerializer,
};
use crate::{
handlers::{block_handler::BlockMessage, peer_handler::models::PeerManagementCmd},
messages::MessagesSerializer,
wrap_network::ActiveConnectionsTrait,
};
use crossbeam::channel::RecvTimeoutError;
use massa_channel::{receiver::MassaReceiver, sender::MassaSender};
use massa_models::block_header::SecuredHeader;
use massa_models::block_id::BlockId;
use massa_protocol_exports::PeerId;
use massa_protocol_exports::{ProtocolConfig, ProtocolError};
use massa_storage::Storage;
use schnellru::{ByLength, LruMap};
use std::thread::JoinHandle;
use std::time::Instant;
use tracing::{debug, info, warn};
// protocol-block-handler-propagation
const THREAD_NAME: &str = "pbh-propagation";
static_assertions::const_assert!(THREAD_NAME.len() < 16);
#[derive(Debug)]
struct BlockPropagationData {
/// Time when propagation was initiated
pub time_added: Instant,
/// Storage holding the block and its dependencies during its propagation time
pub _storage: Storage,
/// Clone of the block header to avoid locking storage during propagation
pub header: SecuredHeader,
}
pub struct PropagationThread {
/// Receiver for commands
receiver: MassaReceiver<BlockHandlerPropagationCommand>,
/// Protocol config
config: ProtocolConfig,
/// Shared access to the block cache
cache: SharedBlockCache,
/// Blocks stored for propagation
stored_for_propagation: LruMap<BlockId, BlockPropagationData>,
/// Shared access to the list of peers connected to us
active_connections: Box<dyn ActiveConnectionsTrait>,
/// Channel to send commands to the peer management system (for banning peers)
peer_cmd_sender: MassaSender<PeerManagementCmd>,
/// Serializer for block-related messages
block_serializer: MessagesSerializer,
}
impl PropagationThread {
fn run(&mut self) {
let tick_interval = self.config.block_propagation_tick.to_duration();
let mut deadline = Instant::now()
.checked_add(tick_interval)
.expect("could not get time of next propagation tick");
loop {
match self.receiver.recv_deadline(deadline) {
Ok(command) => {
match command {
// Message: the block was integrated and should be propagated
BlockHandlerPropagationCommand::IntegratedBlock { block_id, storage } => {
debug!("received IntegratedBlock({})", block_id);
// get the block header
let header = match storage
.read_blocks()
.get(&block_id)
.map(|block| block.content.header.clone())
{
Some(h) => h,
None => {
warn!(
"claimed block {} absent from storage on propagation",
block_id
);
continue;
}
};
// Add the block and its dependencies to the propagation LRU
// to ensure they are stored for the time of the propagation.
self.stored_for_propagation.insert(
block_id,
BlockPropagationData {
time_added: Instant::now(),
_storage: storage,
header,
},
);
// propagate everything that needs to be propagated
self.perform_propagations();
// renew tick because propagation propagations were updated
deadline = Instant::now()
.checked_add(tick_interval)
.expect("could not get time of next propagation tick");
}
BlockHandlerPropagationCommand::AttackBlockDetected(block_id) => {
debug!("received AttackBlockDetected({})", block_id);
let peers_to_ban: Vec<PeerId> = self
.cache
.read()
.blocks_known_by_peer
.iter()
.filter_map(|(peer_id, knowledge)| {
match knowledge.peek(&block_id) {
Some((true, _)) => Some(*peer_id),
_ => None,
}
})
.collect();
self.ban_peers(&peers_to_ban);
}
BlockHandlerPropagationCommand::Stop => {
info!("Stop block propagation thread");
return;
}
}
}
Err(RecvTimeoutError::Timeout) => {
// Propagation tick. This is useful to quickly propagate headers to newly connected nodes.
self.perform_propagations();
// renew deadline of next tick
deadline = Instant::now()
.checked_add(tick_interval)
.expect("could not get time of next propagation tick");
}
Err(RecvTimeoutError::Disconnected) => {
info!("Stop block propagation thread");
return;
}
}
}
}
/// Propagate blocks to peers that need them
fn perform_propagations(&mut self) {
let now = Instant::now();
// stop propagating blocks that have been propagating for too long
while let Some(time_added) = self
.stored_for_propagation
.peek_oldest()
.map(|(_, BlockPropagationData { time_added, .. })| *time_added)
{
if now.saturating_duration_since(time_added)
> self.config.max_block_propagation_time.to_duration()
{
self.stored_for_propagation.pop_oldest();
} else {
break;
}
}
// update caches based on currently connected peers
let peers_connected = self.active_connections.get_peer_ids_connected();
let mut cache_lock = self.cache.write();
cache_lock.update_cache(&peers_connected);
'peer_loop: for (peer_id, known_by_peer) in cache_lock.blocks_known_by_peer.iter_mut() {
for (block_id, BlockPropagationData { header, .. }) in
self.stored_for_propagation.iter()
{
// if the peer already knows about the block, do not propagate it
if let Some((true, _)) = known_by_peer.peek(block_id) {
continue;
}
// try to propagate
debug!("announcing header {} to peer {}", block_id, peer_id);
match self.active_connections.send_to_peer(
peer_id,
&self.block_serializer,
BlockMessage::Header(header.clone()).into(),
true,
) {
Ok(()) => {
// mark the block as known by the peer
known_by_peer.insert(*block_id, (true, now));
}
Err(err) => {
warn!(
"Error while announcing block header {} to peer {} err: {:?}",
block_id, peer_id, err
);
continue 'peer_loop; // try next peer
}
}
}
}
}
/// try to ban a list of peers
fn ban_peers(&mut self, peer_ids: &[PeerId]) {
if let Err(err) = self
.peer_cmd_sender
.try_send(PeerManagementCmd::Ban(peer_ids.to_vec()))
.map_err(|err| ProtocolError::SendError(err.to_string()))
{
warn!("could not send Ban command to peer manager: {}", err);
}
}
}
pub fn start_propagation_thread(
active_connections: Box<dyn ActiveConnectionsTrait>,
receiver: MassaReceiver<BlockHandlerPropagationCommand>,
peer_cmd_sender: MassaSender<PeerManagementCmd>,
config: ProtocolConfig,
cache: SharedBlockCache,
) -> JoinHandle<()> {
std::thread::Builder::new()
.name(THREAD_NAME.to_string())
.spawn(move || {
let block_serializer = MessagesSerializer::new()
.with_block_message_serializer(BlockMessageSerializer::new());
let mut propagation_thread = PropagationThread {
stored_for_propagation: LruMap::new(ByLength::new(
config
.max_blocks_kept_for_propagation
.try_into()
.expect("max_blocks_kept_for_propagation does not fit in u32"),
)),
receiver,
config,
cache,
peer_cmd_sender,
active_connections,
block_serializer,
};
propagation_thread.run();
})
.expect("OS failed to start block propagation thread")
}