rust-libp2p/swarm/src/behaviour.rs
Roman Borschel 8337687b3a
Multiple connections per peer (#1440)
* Allow multiple connections per peer in libp2p-core.

Instead of trying to enforce a single connection per peer,
which involves quite a bit of additional complexity e.g.
to prioritise simultaneously opened connections and can
have other undesirable consequences [1], we now
make multiple connections per peer a feature.

The gist of these changes is as follows:

The concept of a "node" with an implicit 1-1 correspondence
to a connection has been replaced with the "first-class"
concept of a "connection". The code from `src/nodes` has moved
(with varying degrees of modification) to `src/connection`.
A `HandledNode` has become a `Connection`, a `NodeHandler` a
`ConnectionHandler`, the `CollectionStream` was the basis for
the new `connection::Pool`, and so forth.

Conceptually, a `Network` contains a `connection::Pool` which
in turn internally employs the `connection::Manager` for
handling the background `connection::manager::Task`s, one
per connection, as before. These are all considered implementation
details. On the public API, `Peer`s are managed as before through
the `Network`, except now the API has changed with the shift of focus
to (potentially multiple) connections per peer. The `NetworkEvent`s have
accordingly also undergone changes.

The Swarm APIs remain largely unchanged, except for the fact that
`inject_replaced` is no longer called. It may now practically happen
that multiple `ProtocolsHandler`s are associated with a single
`NetworkBehaviour`, one per connection. If implementations of
`NetworkBehaviour` rely somehow on communicating with exactly
one `ProtocolsHandler`, this may cause issues, but it is unlikely.

[1]: https://github.com/paritytech/substrate/issues/4272

* Fix intra-rustdoc links.

* Update core/src/connection/pool.rs

Co-Authored-By: Max Inden <mail@max-inden.de>

* Address some review feedback and fix doc links.

* Allow responses to be sent on the same connection.

* Remove unnecessary remainders of inject_replaced.

* Update swarm/src/behaviour.rs

Co-Authored-By: Pierre Krieger <pierre.krieger1708@gmail.com>

* Update swarm/src/lib.rs

Co-Authored-By: Pierre Krieger <pierre.krieger1708@gmail.com>

* Update core/src/connection/manager.rs

Co-Authored-By: Pierre Krieger <pierre.krieger1708@gmail.com>

* Update core/src/connection/manager.rs

Co-Authored-By: Pierre Krieger <pierre.krieger1708@gmail.com>

* Update core/src/connection/pool.rs

Co-Authored-By: Pierre Krieger <pierre.krieger1708@gmail.com>

* Incorporate more review feedback.

* Move module declaration below imports.

* Update core/src/connection/manager.rs

Co-Authored-By: Toralf Wittner <tw@dtex.org>

* Update core/src/connection/manager.rs

Co-Authored-By: Toralf Wittner <tw@dtex.org>

* Simplify as per review.

* Fix rustoc link.

* Add try_notify_handler and simplify.

* Relocate DialingConnection and DialingAttempt.

For better visibility constraints.

* Small cleanup.

* Small cleanup. More robust EstablishedConnectionIter.

* Clarify semantics of `DialingPeer::connect`.

* Don't call inject_disconnected on InvalidPeerId.

To preserve the previous behavior and ensure calls to
`inject_disconnected` are always paired with calls to
`inject_connected`.

* Provide public ConnectionId constructor.

Mainly needed for testing purposes, e.g. in substrate.

* Move the established connection limit check to the right place.

* Clean up connection error handling.

Separate connection errors into those occuring during
connection setup or upon rejecting a newly established
connection (the `PendingConnectionError`) and those
errors occurring on previously established connections,
i.e. for which a `ConnectionEstablished` event has
been emitted by the connection pool earlier.

* Revert change in log level and clarify an invariant.

* Remove inject_replaced entirely.

* Allow notifying all connection handlers.

Thereby simplify by introducing a new enum `NotifyHandler`,
used with a single constructor `NetworkBehaviourAction::NotifyHandler`.

* Finishing touches.

Small API simplifications and code deduplication.
Some more useful debug logging.

Co-authored-by: Max Inden <mail@max-inden.de>
Co-authored-by: Pierre Krieger <pierre.krieger1708@gmail.com>
Co-authored-by: Toralf Wittner <tw@dtex.org>
2020-03-04 13:49:25 +01:00

256 lines
12 KiB
Rust

// Copyright 2019 Parity Technologies (UK) Ltd.
//
// Permission is hereby granted, free of charge, to any person obtaining a
// copy of this software and associated documentation files (the "Software"),
// to deal in the Software without restriction, including without limitation
// the rights to use, copy, modify, merge, publish, distribute, sublicense,
// and/or sell copies of the Software, and to permit persons to whom the
// Software is furnished to do so, subject to the following conditions:
//
// The above copyright notice and this permission notice shall be included in
// all copies or substantial portions of the Software.
//
// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
// OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
// FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
// DEALINGS IN THE SOFTWARE.
use crate::protocols_handler::{IntoProtocolsHandler, ProtocolsHandler};
use libp2p_core::{ConnectedPoint, Multiaddr, PeerId, connection::{ConnectionId, ListenerId}};
use std::{error, task::Context, task::Poll};
/// A behaviour for the network. Allows customizing the swarm.
///
/// This trait has been designed to be composable. Multiple implementations can be combined into
/// one that handles all the behaviours at once.
///
/// # Deriving `NetworkBehaviour`
///
/// Crate users can implement this trait by using the the `#[derive(NetworkBehaviour)]`
/// proc macro re-exported by the `libp2p` crate. The macro generates a delegating `trait`
/// implementation for the `struct`, which delegates method calls to all trait members. Any events
/// generated by struct members are delegated to [`NetworkBehaviourEventProcess`] implementations
/// which are expected to be provided by the user.
///
/// Optionally one can implement a custom `poll` function, which needs to be tagged with the
/// `#[behaviour(poll_method = "poll")]` attribute, and would be called last with no parameters.
///
/// By default the derive sets the `NetworkBehaviour::OutEvent` as `()` but this can be overriden
/// with `#[behaviour(out_event = "AnotherType")]`.
///
/// `#[behaviour(ignore)]` can be added on a struct field to disable generation of delegation to
/// the fields which do not implement `NetworkBehaviour`.
pub trait NetworkBehaviour: Send + 'static {
/// Handler for all the protocols the network behaviour supports.
type ProtocolsHandler: IntoProtocolsHandler;
/// Event generated by the `NetworkBehaviour` and that the swarm will report back.
type OutEvent: Send + 'static;
/// Creates a new `ProtocolsHandler` for a connection with a peer.
///
/// Every time an incoming connection is opened, and every time we start dialing a node, this
/// method is called.
///
/// The returned object is a handler for that specific connection, and will be moved to a
/// background task dedicated to that connection.
///
/// The network behaviour (ie. the implementation of this trait) and the handlers it has
/// spawned (ie. the objects returned by `new_handler`) can communicate by passing messages.
/// Messages sent from the handler to the behaviour are injected with `inject_event`, and
/// the behaviour can send a message to the handler by making `poll` return `SendEvent`.
fn new_handler(&mut self) -> Self::ProtocolsHandler;
/// Addresses that this behaviour is aware of for this specific peer, and that may allow
/// reaching the peer.
///
/// The addresses will be tried in the order returned by this function, which means that they
/// should be ordered by decreasing likelihood of reachability. In other words, the first
/// address should be the most likely to be reachable.
fn addresses_of_peer(&mut self, peer_id: &PeerId) -> Vec<Multiaddr>;
/// Indicates the behaviour that we connected to the node with the given peer id through the
/// given endpoint.
///
/// This node now has a handler (as spawned by `new_handler`) running in the background.
fn inject_connected(&mut self, peer_id: PeerId, endpoint: ConnectedPoint);
/// Indicates the behaviour that we disconnected from the node with the given peer id. The
/// endpoint is the one we used to be connected to.
///
/// There is no handler running anymore for this node. Any event that has been sent to it may
/// or may not have been processed by the handler.
fn inject_disconnected(&mut self, peer_id: &PeerId, endpoint: ConnectedPoint);
/// Informs the behaviour about an event generated by the handler dedicated to the peer identified by `peer_id`.
/// for the behaviour.
///
/// The `peer_id` is guaranteed to be in a connected state. In other words, `inject_connected`
/// has previously been called with this `PeerId`.
fn inject_event(
&mut self,
peer_id: PeerId,
connection: ConnectionId,
event: <<Self::ProtocolsHandler as IntoProtocolsHandler>::Handler as ProtocolsHandler>::OutEvent
);
/// Indicates to the behaviour that we tried to reach an address, but failed.
///
/// If we were trying to reach a specific node, its ID is passed as parameter. If this is the
/// last address to attempt for the given node, then `inject_dial_failure` is called afterwards.
fn inject_addr_reach_failure(&mut self, _peer_id: Option<&PeerId>, _addr: &Multiaddr, _error: &dyn error::Error) {
}
/// Indicates to the behaviour that we tried to dial all the addresses known for a node, but
/// failed.
///
/// The `peer_id` is guaranteed to be in a disconnected state. In other words,
/// `inject_connected` has not been called, or `inject_disconnected` has been called since then.
fn inject_dial_failure(&mut self, _peer_id: &PeerId) {
}
/// Indicates to the behaviour that we have started listening on a new multiaddr.
fn inject_new_listen_addr(&mut self, _addr: &Multiaddr) {
}
/// Indicates to the behaviour that a new multiaddr we were listening on has expired,
/// which means that we are no longer listening in it.
fn inject_expired_listen_addr(&mut self, _addr: &Multiaddr) {
}
/// Indicates to the behaviour that we have discovered a new external address for us.
fn inject_new_external_addr(&mut self, _addr: &Multiaddr) {
}
/// A listener experienced an error.
fn inject_listener_error(&mut self, _id: ListenerId, _err: &(dyn std::error::Error + 'static)) {
}
/// A listener closed.
fn inject_listener_closed(&mut self, _id: ListenerId) {
}
/// Polls for things that swarm should do.
///
/// This API mimics the API of the `Stream` trait. The method may register the current task in
/// order to wake it up at a later point in time.
fn poll(&mut self, cx: &mut Context, params: &mut impl PollParameters)
-> Poll<NetworkBehaviourAction<<<Self::ProtocolsHandler as IntoProtocolsHandler>::Handler as ProtocolsHandler>::InEvent, Self::OutEvent>>;
}
/// Parameters passed to `poll()`, that the `NetworkBehaviour` has access to.
pub trait PollParameters {
/// Iterator returned by [`supported_protocols`](PollParameters::supported_protocols).
type SupportedProtocolsIter: ExactSizeIterator<Item = Vec<u8>>;
/// Iterator returned by [`listened_addresses`](PollParameters::listened_addresses).
type ListenedAddressesIter: ExactSizeIterator<Item = Multiaddr>;
/// Iterator returned by [`external_addresses`](PollParameters::external_addresses).
type ExternalAddressesIter: ExactSizeIterator<Item = Multiaddr>;
/// Returns the list of protocol the behaviour supports when a remote negotiates a protocol on
/// an inbound substream.
///
/// The iterator's elements are the ASCII names as reported on the wire.
///
/// Note that the list is computed once at initialization and never refreshed.
fn supported_protocols(&self) -> Self::SupportedProtocolsIter;
/// Returns the list of the addresses we're listening on.
fn listened_addresses(&self) -> Self::ListenedAddressesIter;
/// Returns the list of the addresses nodes can use to reach us.
fn external_addresses(&self) -> Self::ExternalAddressesIter;
/// Returns the peer id of the local node.
fn local_peer_id(&self) -> &PeerId;
}
/// When deriving [`NetworkBehaviour`] this trait must be implemented for all the possible event types
/// generated by the inner behaviours.
pub trait NetworkBehaviourEventProcess<TEvent> {
/// Called when one of the fields of the type you're deriving `NetworkBehaviour` on generates
/// an event.
fn inject_event(&mut self, event: TEvent);
}
/// An action that a [`NetworkBehaviour`] can trigger in the [`Swarm`]
/// in whose context it is executing.
///
/// [`Swarm`]: super::Swarm
#[derive(Debug, Clone)]
pub enum NetworkBehaviourAction<TInEvent, TOutEvent> {
/// Instructs the `Swarm` to return an event when it is being polled.
GenerateEvent(TOutEvent),
/// Instructs the swarm to dial the given multiaddress, with no knowledge of the `PeerId` that
/// may be reached.
DialAddress {
/// The address to dial.
address: Multiaddr,
},
/// Instructs the swarm to dial a known `PeerId`.
///
/// The `addresses_of_peer` method is called to determine which addresses to attempt to reach.
///
/// If we were already trying to dial this node, the addresses that are not yet in the queue of
/// addresses to try are added back to this queue.
///
/// On success, [`NetworkBehaviour::inject_connected`] is invoked.
/// On failure, [`NetworkBehaviour::inject_dial_failure`] is invoked.
DialPeer {
/// The peer to try reach.
peer_id: PeerId,
},
/// Instructs the `Swarm` to send an event to the handler dedicated to a
/// connection with a peer.
///
/// If the `Swarm` is connected to the peer, the message is delivered to the
/// `ProtocolsHandler` instance identified by the peer ID and connection ID.
///
/// If the specified connection no longer exists, the event is silently dropped.
///
/// Typically the connection ID given is the same as the one passed to
/// [`NetworkBehaviour::inject_event`], i.e. whenever the behaviour wishes to
/// respond to a request on the same connection (and possibly the same
/// substream, as per the implementation of `ProtocolsHandler`).
///
/// Note that even if the peer is currently connected, connections can get closed
/// at any time and thus the event may not reach a handler.
NotifyHandler {
/// The peer for whom a `ProtocolsHandler` should be notified.
peer_id: PeerId,
/// The ID of the connection whose `ProtocolsHandler` to notify.
handler: NotifyHandler,
/// The event to send.
event: TInEvent,
},
/// Informs the `Swarm` about a multi-address observed by a remote for
/// the local node.
///
/// It is advisable to issue `ReportObservedAddr` actions at a fixed frequency
/// per node. This way address information will be more accurate over time
/// and individual outliers carry less weight.
ReportObservedAddr {
/// The observed address of the local node.
address: Multiaddr,
},
}
/// The options w.r.t. which connection handlers to notify of an event.
#[derive(Debug, Clone)]
pub enum NotifyHandler {
/// Notify a particular connection handler.
One(ConnectionId),
/// Notify an arbitrary connection handler.
Any,
/// Notify all connection handlers.
All
}