bsc/p2p/server_test.go

632 lines
17 KiB
Go
Raw Normal View History

2015-07-07 03:54:22 +03:00
// Copyright 2014 The go-ethereum Authors
// This file is part of the go-ethereum library.
2015-07-07 03:54:22 +03:00
//
// The go-ethereum library is free software: you can redistribute it and/or modify
2015-07-07 03:54:22 +03:00
// it under the terms of the GNU Lesser General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// The go-ethereum library is distributed in the hope that it will be useful,
2015-07-07 03:54:22 +03:00
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
2015-07-07 03:54:22 +03:00
// GNU Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public License
// along with the go-ethereum library. If not, see <http://www.gnu.org/licenses/>.
2015-07-07 03:54:22 +03:00
2014-10-23 18:57:54 +03:00
package p2p
import (
"crypto/ecdsa"
"crypto/sha256"
"errors"
"io"
"math/rand"
2014-10-23 18:57:54 +03:00
"net"
"reflect"
"strconv"
"strings"
2014-10-23 18:57:54 +03:00
"testing"
"time"
"github.com/ethereum/go-ethereum/crypto"
"github.com/ethereum/go-ethereum/internal/testlog"
"github.com/ethereum/go-ethereum/log"
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
"github.com/ethereum/go-ethereum/p2p/enode"
"github.com/ethereum/go-ethereum/p2p/enr"
"github.com/ethereum/go-ethereum/p2p/rlpx"
2014-10-23 18:57:54 +03:00
)
type testTransport struct {
*rlpxTransport
rpub *ecdsa.PublicKey
closeErr error
}
func newTestTransport(rpub *ecdsa.PublicKey, fd net.Conn, dialDest *ecdsa.PublicKey) transport {
wrapped := newRLPX(fd, dialDest).(*rlpxTransport)
wrapped.conn.InitWithSecrets(rlpx.Secrets{
AES: make([]byte, 16),
MAC: make([]byte, 16),
EgressMAC: sha256.New(),
IngressMAC: sha256.New(),
})
return &testTransport{rpub: rpub, rlpxTransport: wrapped}
}
func (c *testTransport) doEncHandshake(prv *ecdsa.PrivateKey) (*ecdsa.PublicKey, error) {
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
return c.rpub, nil
}
func (c *testTransport) doProtoHandshake(our *protoHandshake) (*protoHandshake, error) {
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
pubkey := crypto.FromECDSAPub(c.rpub)[1:]
return &protoHandshake{ID: pubkey, Name: "test"}, nil
}
func (c *testTransport) close(err error) {
c.conn.Close()
c.closeErr = err
}
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
func startTestServer(t *testing.T, remoteKey *ecdsa.PublicKey, pf func(*Peer)) *Server {
config := Config{
p2p: new dial scheduler (#20592) * p2p: new dial scheduler This change replaces the peer-to-peer dial scheduler with a new and improved implementation. The new code is better than the previous implementation in two key aspects: - The time between discovery of a node and dialing that node is significantly lower in the new version. The old dialState kept a buffer of nodes and launched a task to refill it whenever the buffer became empty. This worked well with the discovery interface we used to have, but doesn't really work with the new iterator-based discovery API. - Selection of static dial candidates (created by Server.AddPeer or through static-nodes.json) performs much better for large amounts of static peers. Connections to static nodes are now limited like dynanic dials and can no longer overstep MaxPeers or the dial ratio. * p2p/simulations/adapters: adapt to new NodeDialer interface * p2p: re-add check for self in checkDial * p2p: remove peersetCh * p2p: allow static dials when discovery is disabled * p2p: add test for dialScheduler.removeStatic * p2p: remove blank line * p2p: fix documentation of maxDialPeers * p2p: change "ok" to "added" in static node log * p2p: improve dialTask docs Also increase log level for "Can't resolve node" * p2p: ensure dial resolver is truly nil without discovery * p2p: add "looking for peers" log message * p2p: clean up Server.run comments * p2p: fix maxDialedConns for maxpeers < dialRatio Always allocate at least one dial slot unless dialing is disabled using NoDial or MaxPeers == 0. Most importantly, this fixes MaxPeers == 1 to dedicate the sole slot to dialing instead of listening. * p2p: fix RemovePeer to disconnect the peer again Also make RemovePeer synchronous and add a test. * p2p: remove "Connection set up" log message * p2p: clean up connection logging We previously logged outgoing connection failures up to three times. - in SetupConn() as "Setting up connection failed addr=..." - in setupConn() with an error-specific message and "id=... addr=..." - in dial() as "Dial error task=..." This commit ensures a single log message is emitted per failure and adds "id=... addr=... conn=..." everywhere (id= omitted when the ID isn't known yet). Also avoid printing a log message when a static dial fails but can't be resolved because discv4 is disabled. The light client hit this case all the time, increasing the message count to four lines per failed connection. * p2p: document that RemovePeer blocks
2020-02-13 13:10:03 +03:00
Name: "test",
MaxPeers: 10,
ListenAddr: "127.0.0.1:0",
NoDiscovery: true,
PrivateKey: newkey(),
Logger: testlog.Logger(t, log.LvlTrace),
}
server := &Server{
Config: config,
newPeerHook: pf,
newTransport: func(fd net.Conn, dialDest *ecdsa.PublicKey) transport {
return newTestTransport(remoteKey, fd, dialDest)
},
2014-10-23 18:57:54 +03:00
}
if err := server.Start(); err != nil {
t.Fatalf("Could not start server: %v", err)
2014-10-23 18:57:54 +03:00
}
return server
2014-10-23 18:57:54 +03:00
}
func TestServerListen(t *testing.T) {
// start the test server
connected := make(chan *Peer)
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
remid := &newkey().PublicKey
srv := startTestServer(t, remid, func(p *Peer) {
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
if p.ID() != enode.PubkeyToIDV4(remid) {
t.Error("peer func called with wrong node id")
}
connected <- p
})
defer close(connected)
defer srv.Stop()
// dial the test server
conn, err := net.DialTimeout("tcp", srv.ListenAddr, 5*time.Second)
if err != nil {
t.Fatalf("could not dial: %v", err)
2014-11-04 14:21:44 +02:00
}
defer conn.Close()
2014-11-04 14:21:44 +02:00
select {
case peer := <-connected:
if peer.LocalAddr().String() != conn.RemoteAddr().String() {
t.Errorf("peer started with wrong conn: got %v, want %v",
peer.LocalAddr(), conn.RemoteAddr())
}
peers := srv.Peers()
if !reflect.DeepEqual(peers, []*Peer{peer}) {
t.Errorf("Peers mismatch: got %v, want %v", peers, []*Peer{peer})
}
case <-time.After(1 * time.Second):
t.Error("server did not accept within one second")
2014-11-04 14:21:44 +02:00
}
}
func TestServerDial(t *testing.T) {
// run a one-shot TCP server to handle the connection.
listener, err := net.Listen("tcp", "127.0.0.1:0")
if err != nil {
2016-04-15 12:06:57 +03:00
t.Fatalf("could not setup listener: %v", err)
}
defer listener.Close()
p2p: new dial scheduler (#20592) * p2p: new dial scheduler This change replaces the peer-to-peer dial scheduler with a new and improved implementation. The new code is better than the previous implementation in two key aspects: - The time between discovery of a node and dialing that node is significantly lower in the new version. The old dialState kept a buffer of nodes and launched a task to refill it whenever the buffer became empty. This worked well with the discovery interface we used to have, but doesn't really work with the new iterator-based discovery API. - Selection of static dial candidates (created by Server.AddPeer or through static-nodes.json) performs much better for large amounts of static peers. Connections to static nodes are now limited like dynanic dials and can no longer overstep MaxPeers or the dial ratio. * p2p/simulations/adapters: adapt to new NodeDialer interface * p2p: re-add check for self in checkDial * p2p: remove peersetCh * p2p: allow static dials when discovery is disabled * p2p: add test for dialScheduler.removeStatic * p2p: remove blank line * p2p: fix documentation of maxDialPeers * p2p: change "ok" to "added" in static node log * p2p: improve dialTask docs Also increase log level for "Can't resolve node" * p2p: ensure dial resolver is truly nil without discovery * p2p: add "looking for peers" log message * p2p: clean up Server.run comments * p2p: fix maxDialedConns for maxpeers < dialRatio Always allocate at least one dial slot unless dialing is disabled using NoDial or MaxPeers == 0. Most importantly, this fixes MaxPeers == 1 to dedicate the sole slot to dialing instead of listening. * p2p: fix RemovePeer to disconnect the peer again Also make RemovePeer synchronous and add a test. * p2p: remove "Connection set up" log message * p2p: clean up connection logging We previously logged outgoing connection failures up to three times. - in SetupConn() as "Setting up connection failed addr=..." - in setupConn() with an error-specific message and "id=... addr=..." - in dial() as "Dial error task=..." This commit ensures a single log message is emitted per failure and adds "id=... addr=... conn=..." everywhere (id= omitted when the ID isn't known yet). Also avoid printing a log message when a static dial fails but can't be resolved because discv4 is disabled. The light client hit this case all the time, increasing the message count to four lines per failed connection. * p2p: document that RemovePeer blocks
2020-02-13 13:10:03 +03:00
accepted := make(chan net.Conn, 1)
go func() {
conn, err := listener.Accept()
if err != nil {
return
}
accepted <- conn
}()
// start the server
connected := make(chan *Peer)
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
remid := &newkey().PublicKey
srv := startTestServer(t, remid, func(p *Peer) { connected <- p })
defer close(connected)
defer srv.Stop()
// tell the server to connect
tcpAddr := listener.Addr().(*net.TCPAddr)
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
node := enode.NewV4(remid, tcpAddr.IP, tcpAddr.Port, 0)
srv.AddPeer(node)
select {
case conn := <-accepted:
2015-06-09 22:26:26 +03:00
defer conn.Close()
2014-11-04 14:21:44 +02:00
select {
case peer := <-connected:
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
if peer.ID() != enode.PubkeyToIDV4(remid) {
t.Errorf("peer has wrong id")
}
if peer.Name() != "test" {
t.Errorf("peer has wrong name")
}
if peer.RemoteAddr().String() != conn.LocalAddr().String() {
t.Errorf("peer started with wrong conn: got %v, want %v",
peer.RemoteAddr(), conn.LocalAddr())
2014-11-04 14:21:44 +02:00
}
peers := srv.Peers()
if !reflect.DeepEqual(peers, []*Peer{peer}) {
t.Errorf("Peers mismatch: got %v, want %v", peers, []*Peer{peer})
}
// Test AddTrustedPeer/RemoveTrustedPeer and changing Trusted flags
// Particularly for race conditions on changing the flag state.
if peer := srv.Peers()[0]; peer.Info().Network.Trusted {
t.Errorf("peer is trusted prematurely: %v", peer)
}
done := make(chan bool)
go func() {
srv.AddTrustedPeer(node)
if peer := srv.Peers()[0]; !peer.Info().Network.Trusted {
t.Errorf("peer is not trusted after AddTrustedPeer: %v", peer)
}
srv.RemoveTrustedPeer(node)
if peer := srv.Peers()[0]; peer.Info().Network.Trusted {
t.Errorf("peer is trusted after RemoveTrustedPeer: %v", peer)
}
done <- true
}()
// Trigger potential race conditions
peer = srv.Peers()[0]
_ = peer.Inbound()
_ = peer.Info()
<-done
case <-time.After(1 * time.Second):
t.Error("server did not launch peer within one second")
2014-11-04 14:21:44 +02:00
}
case <-time.After(1 * time.Second):
t.Error("server did not connect within one second")
2014-10-23 18:57:54 +03:00
}
}
p2p: new dial scheduler (#20592) * p2p: new dial scheduler This change replaces the peer-to-peer dial scheduler with a new and improved implementation. The new code is better than the previous implementation in two key aspects: - The time between discovery of a node and dialing that node is significantly lower in the new version. The old dialState kept a buffer of nodes and launched a task to refill it whenever the buffer became empty. This worked well with the discovery interface we used to have, but doesn't really work with the new iterator-based discovery API. - Selection of static dial candidates (created by Server.AddPeer or through static-nodes.json) performs much better for large amounts of static peers. Connections to static nodes are now limited like dynanic dials and can no longer overstep MaxPeers or the dial ratio. * p2p/simulations/adapters: adapt to new NodeDialer interface * p2p: re-add check for self in checkDial * p2p: remove peersetCh * p2p: allow static dials when discovery is disabled * p2p: add test for dialScheduler.removeStatic * p2p: remove blank line * p2p: fix documentation of maxDialPeers * p2p: change "ok" to "added" in static node log * p2p: improve dialTask docs Also increase log level for "Can't resolve node" * p2p: ensure dial resolver is truly nil without discovery * p2p: add "looking for peers" log message * p2p: clean up Server.run comments * p2p: fix maxDialedConns for maxpeers < dialRatio Always allocate at least one dial slot unless dialing is disabled using NoDial or MaxPeers == 0. Most importantly, this fixes MaxPeers == 1 to dedicate the sole slot to dialing instead of listening. * p2p: fix RemovePeer to disconnect the peer again Also make RemovePeer synchronous and add a test. * p2p: remove "Connection set up" log message * p2p: clean up connection logging We previously logged outgoing connection failures up to three times. - in SetupConn() as "Setting up connection failed addr=..." - in setupConn() with an error-specific message and "id=... addr=..." - in dial() as "Dial error task=..." This commit ensures a single log message is emitted per failure and adds "id=... addr=... conn=..." everywhere (id= omitted when the ID isn't known yet). Also avoid printing a log message when a static dial fails but can't be resolved because discv4 is disabled. The light client hit this case all the time, increasing the message count to four lines per failed connection. * p2p: document that RemovePeer blocks
2020-02-13 13:10:03 +03:00
// This test checks that RemovePeer disconnects the peer if it is connected.
func TestServerRemovePeerDisconnect(t *testing.T) {
srv1 := &Server{Config: Config{
PrivateKey: newkey(),
MaxPeers: 1,
NoDiscovery: true,
Logger: testlog.Logger(t, log.LvlTrace).New("server", "1"),
}}
srv2 := &Server{Config: Config{
PrivateKey: newkey(),
MaxPeers: 1,
NoDiscovery: true,
NoDial: true,
ListenAddr: "127.0.0.1:0",
Logger: testlog.Logger(t, log.LvlTrace).New("server", "2"),
}}
srv1.Start()
defer srv1.Stop()
srv2.Start()
defer srv2.Stop()
s := strings.Split(srv2.ListenAddr, ":")
if len(s) != 2 {
t.Fatal("invalid ListenAddr")
}
if port, err := strconv.Atoi(s[1]); err == nil {
srv2.localnode.Set(enr.TCP(uint16(port)))
}
p2p: new dial scheduler (#20592) * p2p: new dial scheduler This change replaces the peer-to-peer dial scheduler with a new and improved implementation. The new code is better than the previous implementation in two key aspects: - The time between discovery of a node and dialing that node is significantly lower in the new version. The old dialState kept a buffer of nodes and launched a task to refill it whenever the buffer became empty. This worked well with the discovery interface we used to have, but doesn't really work with the new iterator-based discovery API. - Selection of static dial candidates (created by Server.AddPeer or through static-nodes.json) performs much better for large amounts of static peers. Connections to static nodes are now limited like dynanic dials and can no longer overstep MaxPeers or the dial ratio. * p2p/simulations/adapters: adapt to new NodeDialer interface * p2p: re-add check for self in checkDial * p2p: remove peersetCh * p2p: allow static dials when discovery is disabled * p2p: add test for dialScheduler.removeStatic * p2p: remove blank line * p2p: fix documentation of maxDialPeers * p2p: change "ok" to "added" in static node log * p2p: improve dialTask docs Also increase log level for "Can't resolve node" * p2p: ensure dial resolver is truly nil without discovery * p2p: add "looking for peers" log message * p2p: clean up Server.run comments * p2p: fix maxDialedConns for maxpeers < dialRatio Always allocate at least one dial slot unless dialing is disabled using NoDial or MaxPeers == 0. Most importantly, this fixes MaxPeers == 1 to dedicate the sole slot to dialing instead of listening. * p2p: fix RemovePeer to disconnect the peer again Also make RemovePeer synchronous and add a test. * p2p: remove "Connection set up" log message * p2p: clean up connection logging We previously logged outgoing connection failures up to three times. - in SetupConn() as "Setting up connection failed addr=..." - in setupConn() with an error-specific message and "id=... addr=..." - in dial() as "Dial error task=..." This commit ensures a single log message is emitted per failure and adds "id=... addr=... conn=..." everywhere (id= omitted when the ID isn't known yet). Also avoid printing a log message when a static dial fails but can't be resolved because discv4 is disabled. The light client hit this case all the time, increasing the message count to four lines per failed connection. * p2p: document that RemovePeer blocks
2020-02-13 13:10:03 +03:00
if !syncAddPeer(srv1, srv2.Self()) {
t.Fatal("peer not connected")
2015-05-04 13:08:42 +03:00
}
p2p: new dial scheduler (#20592) * p2p: new dial scheduler This change replaces the peer-to-peer dial scheduler with a new and improved implementation. The new code is better than the previous implementation in two key aspects: - The time between discovery of a node and dialing that node is significantly lower in the new version. The old dialState kept a buffer of nodes and launched a task to refill it whenever the buffer became empty. This worked well with the discovery interface we used to have, but doesn't really work with the new iterator-based discovery API. - Selection of static dial candidates (created by Server.AddPeer or through static-nodes.json) performs much better for large amounts of static peers. Connections to static nodes are now limited like dynanic dials and can no longer overstep MaxPeers or the dial ratio. * p2p/simulations/adapters: adapt to new NodeDialer interface * p2p: re-add check for self in checkDial * p2p: remove peersetCh * p2p: allow static dials when discovery is disabled * p2p: add test for dialScheduler.removeStatic * p2p: remove blank line * p2p: fix documentation of maxDialPeers * p2p: change "ok" to "added" in static node log * p2p: improve dialTask docs Also increase log level for "Can't resolve node" * p2p: ensure dial resolver is truly nil without discovery * p2p: add "looking for peers" log message * p2p: clean up Server.run comments * p2p: fix maxDialedConns for maxpeers < dialRatio Always allocate at least one dial slot unless dialing is disabled using NoDial or MaxPeers == 0. Most importantly, this fixes MaxPeers == 1 to dedicate the sole slot to dialing instead of listening. * p2p: fix RemovePeer to disconnect the peer again Also make RemovePeer synchronous and add a test. * p2p: remove "Connection set up" log message * p2p: clean up connection logging We previously logged outgoing connection failures up to three times. - in SetupConn() as "Setting up connection failed addr=..." - in setupConn() with an error-specific message and "id=... addr=..." - in dial() as "Dial error task=..." This commit ensures a single log message is emitted per failure and adds "id=... addr=... conn=..." everywhere (id= omitted when the ID isn't known yet). Also avoid printing a log message when a static dial fails but can't be resolved because discv4 is disabled. The light client hit this case all the time, increasing the message count to four lines per failed connection. * p2p: document that RemovePeer blocks
2020-02-13 13:10:03 +03:00
srv1.RemovePeer(srv2.Self())
if srv1.PeerCount() > 0 {
t.Fatal("removed peer still connected")
2015-05-04 13:08:42 +03:00
}
}
2015-05-04 13:08:42 +03:00
p2p: new dial scheduler (#20592) * p2p: new dial scheduler This change replaces the peer-to-peer dial scheduler with a new and improved implementation. The new code is better than the previous implementation in two key aspects: - The time between discovery of a node and dialing that node is significantly lower in the new version. The old dialState kept a buffer of nodes and launched a task to refill it whenever the buffer became empty. This worked well with the discovery interface we used to have, but doesn't really work with the new iterator-based discovery API. - Selection of static dial candidates (created by Server.AddPeer or through static-nodes.json) performs much better for large amounts of static peers. Connections to static nodes are now limited like dynanic dials and can no longer overstep MaxPeers or the dial ratio. * p2p/simulations/adapters: adapt to new NodeDialer interface * p2p: re-add check for self in checkDial * p2p: remove peersetCh * p2p: allow static dials when discovery is disabled * p2p: add test for dialScheduler.removeStatic * p2p: remove blank line * p2p: fix documentation of maxDialPeers * p2p: change "ok" to "added" in static node log * p2p: improve dialTask docs Also increase log level for "Can't resolve node" * p2p: ensure dial resolver is truly nil without discovery * p2p: add "looking for peers" log message * p2p: clean up Server.run comments * p2p: fix maxDialedConns for maxpeers < dialRatio Always allocate at least one dial slot unless dialing is disabled using NoDial or MaxPeers == 0. Most importantly, this fixes MaxPeers == 1 to dedicate the sole slot to dialing instead of listening. * p2p: fix RemovePeer to disconnect the peer again Also make RemovePeer synchronous and add a test. * p2p: remove "Connection set up" log message * p2p: clean up connection logging We previously logged outgoing connection failures up to three times. - in SetupConn() as "Setting up connection failed addr=..." - in setupConn() with an error-specific message and "id=... addr=..." - in dial() as "Dial error task=..." This commit ensures a single log message is emitted per failure and adds "id=... addr=... conn=..." everywhere (id= omitted when the ID isn't known yet). Also avoid printing a log message when a static dial fails but can't be resolved because discv4 is disabled. The light client hit this case all the time, increasing the message count to four lines per failed connection. * p2p: document that RemovePeer blocks
2020-02-13 13:10:03 +03:00
// This test checks that connections are disconnected just after the encryption handshake
// when the server is at capacity. Trusted connections should still be accepted.
func TestServerAtCap(t *testing.T) {
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
trustedNode := newkey()
trustedID := enode.PubkeyToIDV4(&trustedNode.PublicKey)
srv := &Server{
Config: Config{
PrivateKey: newkey(),
MaxPeers: 10,
NoDial: true,
NoDiscovery: true,
p2p: new dial scheduler (#20592) * p2p: new dial scheduler This change replaces the peer-to-peer dial scheduler with a new and improved implementation. The new code is better than the previous implementation in two key aspects: - The time between discovery of a node and dialing that node is significantly lower in the new version. The old dialState kept a buffer of nodes and launched a task to refill it whenever the buffer became empty. This worked well with the discovery interface we used to have, but doesn't really work with the new iterator-based discovery API. - Selection of static dial candidates (created by Server.AddPeer or through static-nodes.json) performs much better for large amounts of static peers. Connections to static nodes are now limited like dynanic dials and can no longer overstep MaxPeers or the dial ratio. * p2p/simulations/adapters: adapt to new NodeDialer interface * p2p: re-add check for self in checkDial * p2p: remove peersetCh * p2p: allow static dials when discovery is disabled * p2p: add test for dialScheduler.removeStatic * p2p: remove blank line * p2p: fix documentation of maxDialPeers * p2p: change "ok" to "added" in static node log * p2p: improve dialTask docs Also increase log level for "Can't resolve node" * p2p: ensure dial resolver is truly nil without discovery * p2p: add "looking for peers" log message * p2p: clean up Server.run comments * p2p: fix maxDialedConns for maxpeers < dialRatio Always allocate at least one dial slot unless dialing is disabled using NoDial or MaxPeers == 0. Most importantly, this fixes MaxPeers == 1 to dedicate the sole slot to dialing instead of listening. * p2p: fix RemovePeer to disconnect the peer again Also make RemovePeer synchronous and add a test. * p2p: remove "Connection set up" log message * p2p: clean up connection logging We previously logged outgoing connection failures up to three times. - in SetupConn() as "Setting up connection failed addr=..." - in setupConn() with an error-specific message and "id=... addr=..." - in dial() as "Dial error task=..." This commit ensures a single log message is emitted per failure and adds "id=... addr=... conn=..." everywhere (id= omitted when the ID isn't known yet). Also avoid printing a log message when a static dial fails but can't be resolved because discv4 is disabled. The light client hit this case all the time, increasing the message count to four lines per failed connection. * p2p: document that RemovePeer blocks
2020-02-13 13:10:03 +03:00
TrustedNodes: []*enode.Node{newNode(trustedID, "")},
Logger: testlog.Logger(t, log.LvlTrace),
},
}
if err := srv.Start(); err != nil {
t.Fatalf("could not start: %v", err)
}
defer srv.Stop()
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
newconn := func(id enode.ID) *conn {
fd, _ := net.Pipe()
tx := newTestTransport(&trustedNode.PublicKey, fd, nil)
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
node := enode.SignNull(new(enr.Record), id)
return &conn{fd: fd, transport: tx, flags: inboundConn, node: node, cont: make(chan error)}
}
// Inject a few connections to fill up the peer set.
for i := 0; i < 10; i++ {
c := newconn(randomID())
if err := srv.checkpoint(c, srv.checkpointAddPeer); err != nil {
t.Fatalf("could not add conn %d: %v", i, err)
}
}
// Try inserting a non-trusted connection.
anotherID := randomID()
c := newconn(anotherID)
if err := srv.checkpoint(c, srv.checkpointPostHandshake); err != DiscTooManyPeers {
t.Error("wrong error for insert:", err)
}
// Try inserting a trusted connection.
c = newconn(trustedID)
if err := srv.checkpoint(c, srv.checkpointPostHandshake); err != nil {
t.Error("unexpected error for trusted conn @posthandshake:", err)
}
if !c.is(trustedConn) {
t.Error("Server did not set trusted flag")
}
// Remove from trusted set and try again
p2p: new dial scheduler (#20592) * p2p: new dial scheduler This change replaces the peer-to-peer dial scheduler with a new and improved implementation. The new code is better than the previous implementation in two key aspects: - The time between discovery of a node and dialing that node is significantly lower in the new version. The old dialState kept a buffer of nodes and launched a task to refill it whenever the buffer became empty. This worked well with the discovery interface we used to have, but doesn't really work with the new iterator-based discovery API. - Selection of static dial candidates (created by Server.AddPeer or through static-nodes.json) performs much better for large amounts of static peers. Connections to static nodes are now limited like dynanic dials and can no longer overstep MaxPeers or the dial ratio. * p2p/simulations/adapters: adapt to new NodeDialer interface * p2p: re-add check for self in checkDial * p2p: remove peersetCh * p2p: allow static dials when discovery is disabled * p2p: add test for dialScheduler.removeStatic * p2p: remove blank line * p2p: fix documentation of maxDialPeers * p2p: change "ok" to "added" in static node log * p2p: improve dialTask docs Also increase log level for "Can't resolve node" * p2p: ensure dial resolver is truly nil without discovery * p2p: add "looking for peers" log message * p2p: clean up Server.run comments * p2p: fix maxDialedConns for maxpeers < dialRatio Always allocate at least one dial slot unless dialing is disabled using NoDial or MaxPeers == 0. Most importantly, this fixes MaxPeers == 1 to dedicate the sole slot to dialing instead of listening. * p2p: fix RemovePeer to disconnect the peer again Also make RemovePeer synchronous and add a test. * p2p: remove "Connection set up" log message * p2p: clean up connection logging We previously logged outgoing connection failures up to three times. - in SetupConn() as "Setting up connection failed addr=..." - in setupConn() with an error-specific message and "id=... addr=..." - in dial() as "Dial error task=..." This commit ensures a single log message is emitted per failure and adds "id=... addr=... conn=..." everywhere (id= omitted when the ID isn't known yet). Also avoid printing a log message when a static dial fails but can't be resolved because discv4 is disabled. The light client hit this case all the time, increasing the message count to four lines per failed connection. * p2p: document that RemovePeer blocks
2020-02-13 13:10:03 +03:00
srv.RemoveTrustedPeer(newNode(trustedID, ""))
c = newconn(trustedID)
if err := srv.checkpoint(c, srv.checkpointPostHandshake); err != DiscTooManyPeers {
t.Error("wrong error for insert:", err)
}
// Add anotherID to trusted set and try again
p2p: new dial scheduler (#20592) * p2p: new dial scheduler This change replaces the peer-to-peer dial scheduler with a new and improved implementation. The new code is better than the previous implementation in two key aspects: - The time between discovery of a node and dialing that node is significantly lower in the new version. The old dialState kept a buffer of nodes and launched a task to refill it whenever the buffer became empty. This worked well with the discovery interface we used to have, but doesn't really work with the new iterator-based discovery API. - Selection of static dial candidates (created by Server.AddPeer or through static-nodes.json) performs much better for large amounts of static peers. Connections to static nodes are now limited like dynanic dials and can no longer overstep MaxPeers or the dial ratio. * p2p/simulations/adapters: adapt to new NodeDialer interface * p2p: re-add check for self in checkDial * p2p: remove peersetCh * p2p: allow static dials when discovery is disabled * p2p: add test for dialScheduler.removeStatic * p2p: remove blank line * p2p: fix documentation of maxDialPeers * p2p: change "ok" to "added" in static node log * p2p: improve dialTask docs Also increase log level for "Can't resolve node" * p2p: ensure dial resolver is truly nil without discovery * p2p: add "looking for peers" log message * p2p: clean up Server.run comments * p2p: fix maxDialedConns for maxpeers < dialRatio Always allocate at least one dial slot unless dialing is disabled using NoDial or MaxPeers == 0. Most importantly, this fixes MaxPeers == 1 to dedicate the sole slot to dialing instead of listening. * p2p: fix RemovePeer to disconnect the peer again Also make RemovePeer synchronous and add a test. * p2p: remove "Connection set up" log message * p2p: clean up connection logging We previously logged outgoing connection failures up to three times. - in SetupConn() as "Setting up connection failed addr=..." - in setupConn() with an error-specific message and "id=... addr=..." - in dial() as "Dial error task=..." This commit ensures a single log message is emitted per failure and adds "id=... addr=... conn=..." everywhere (id= omitted when the ID isn't known yet). Also avoid printing a log message when a static dial fails but can't be resolved because discv4 is disabled. The light client hit this case all the time, increasing the message count to four lines per failed connection. * p2p: document that RemovePeer blocks
2020-02-13 13:10:03 +03:00
srv.AddTrustedPeer(newNode(anotherID, ""))
c = newconn(anotherID)
if err := srv.checkpoint(c, srv.checkpointPostHandshake); err != nil {
t.Error("unexpected error for trusted conn @posthandshake:", err)
}
if !c.is(trustedConn) {
t.Error("Server did not set trusted flag")
}
}
func TestServerPeerLimits(t *testing.T) {
srvkey := newkey()
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
clientkey := newkey()
clientnode := enode.NewV4(&clientkey.PublicKey, nil, 0, 0)
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
var tp = &setupTransport{
pubkey: &clientkey.PublicKey,
phs: protoHandshake{
ID: crypto.FromECDSAPub(&clientkey.PublicKey)[1:],
// Force "DiscUselessPeer" due to unmatching caps
// Caps: []Cap{discard.cap()},
},
}
srv := &Server{
Config: Config{
PrivateKey: srvkey,
MaxPeers: 0,
NoDial: true,
NoDiscovery: true,
Protocols: []Protocol{discard},
p2p: new dial scheduler (#20592) * p2p: new dial scheduler This change replaces the peer-to-peer dial scheduler with a new and improved implementation. The new code is better than the previous implementation in two key aspects: - The time between discovery of a node and dialing that node is significantly lower in the new version. The old dialState kept a buffer of nodes and launched a task to refill it whenever the buffer became empty. This worked well with the discovery interface we used to have, but doesn't really work with the new iterator-based discovery API. - Selection of static dial candidates (created by Server.AddPeer or through static-nodes.json) performs much better for large amounts of static peers. Connections to static nodes are now limited like dynanic dials and can no longer overstep MaxPeers or the dial ratio. * p2p/simulations/adapters: adapt to new NodeDialer interface * p2p: re-add check for self in checkDial * p2p: remove peersetCh * p2p: allow static dials when discovery is disabled * p2p: add test for dialScheduler.removeStatic * p2p: remove blank line * p2p: fix documentation of maxDialPeers * p2p: change "ok" to "added" in static node log * p2p: improve dialTask docs Also increase log level for "Can't resolve node" * p2p: ensure dial resolver is truly nil without discovery * p2p: add "looking for peers" log message * p2p: clean up Server.run comments * p2p: fix maxDialedConns for maxpeers < dialRatio Always allocate at least one dial slot unless dialing is disabled using NoDial or MaxPeers == 0. Most importantly, this fixes MaxPeers == 1 to dedicate the sole slot to dialing instead of listening. * p2p: fix RemovePeer to disconnect the peer again Also make RemovePeer synchronous and add a test. * p2p: remove "Connection set up" log message * p2p: clean up connection logging We previously logged outgoing connection failures up to three times. - in SetupConn() as "Setting up connection failed addr=..." - in setupConn() with an error-specific message and "id=... addr=..." - in dial() as "Dial error task=..." This commit ensures a single log message is emitted per failure and adds "id=... addr=... conn=..." everywhere (id= omitted when the ID isn't known yet). Also avoid printing a log message when a static dial fails but can't be resolved because discv4 is disabled. The light client hit this case all the time, increasing the message count to four lines per failed connection. * p2p: document that RemovePeer blocks
2020-02-13 13:10:03 +03:00
Logger: testlog.Logger(t, log.LvlTrace),
},
newTransport: func(fd net.Conn, dialDest *ecdsa.PublicKey) transport { return tp },
}
if err := srv.Start(); err != nil {
t.Fatalf("couldn't start server: %v", err)
}
defer srv.Stop()
// Check that server is full (MaxPeers=0)
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
flags := dynDialedConn
dialDest := clientnode
conn, _ := net.Pipe()
srv.SetupConn(conn, flags, dialDest)
if tp.closeErr != DiscTooManyPeers {
t.Errorf("unexpected close error: %q", tp.closeErr)
}
conn.Close()
srv.AddTrustedPeer(clientnode)
// Check that server allows a trusted peer despite being full.
conn, _ = net.Pipe()
srv.SetupConn(conn, flags, dialDest)
if tp.closeErr == DiscTooManyPeers {
t.Errorf("failed to bypass MaxPeers with trusted node: %q", tp.closeErr)
}
if tp.closeErr != DiscUselessPeer {
t.Errorf("unexpected close error: %q", tp.closeErr)
}
conn.Close()
srv.RemoveTrustedPeer(clientnode)
// Check that server is full again.
conn, _ = net.Pipe()
srv.SetupConn(conn, flags, dialDest)
if tp.closeErr != DiscTooManyPeers {
t.Errorf("unexpected close error: %q", tp.closeErr)
}
conn.Close()
}
func TestServerSetupConn(t *testing.T) {
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
var (
clientkey, srvkey = newkey(), newkey()
clientpub = &clientkey.PublicKey
srvpub = &srvkey.PublicKey
)
tests := []struct {
dontstart bool
tt *setupTransport
flags connFlag
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
dialDest *enode.Node
wantCloseErr error
wantCalls string
}{
{
dontstart: true,
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
tt: &setupTransport{pubkey: clientpub},
wantCalls: "close,",
wantCloseErr: errServerStopped,
},
{
tt: &setupTransport{pubkey: clientpub, encHandshakeErr: errEncHandshakeError},
flags: inboundConn,
wantCalls: "doEncHandshake,close,",
wantCloseErr: errEncHandshakeError,
},
{
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
tt: &setupTransport{pubkey: clientpub, phs: protoHandshake{ID: randomID().Bytes()}},
dialDest: enode.NewV4(clientpub, nil, 0, 0),
flags: dynDialedConn,
wantCalls: "doEncHandshake,doProtoHandshake,close,",
wantCloseErr: DiscUnexpectedIdentity,
},
{
tt: &setupTransport{pubkey: clientpub, protoHandshakeErr: errProtoHandshakeError},
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
dialDest: enode.NewV4(clientpub, nil, 0, 0),
flags: dynDialedConn,
wantCalls: "doEncHandshake,doProtoHandshake,close,",
wantCloseErr: errProtoHandshakeError,
},
{
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
tt: &setupTransport{pubkey: srvpub, phs: protoHandshake{ID: crypto.FromECDSAPub(srvpub)[1:]}},
flags: inboundConn,
wantCalls: "doEncHandshake,close,",
wantCloseErr: DiscSelf,
},
{
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
tt: &setupTransport{pubkey: clientpub, phs: protoHandshake{ID: crypto.FromECDSAPub(clientpub)[1:]}},
flags: inboundConn,
wantCalls: "doEncHandshake,doProtoHandshake,close,",
wantCloseErr: DiscUselessPeer,
},
}
for i, test := range tests {
t.Run(test.wantCalls, func(t *testing.T) {
cfg := Config{
PrivateKey: srvkey,
MaxPeers: 10,
NoDial: true,
NoDiscovery: true,
Protocols: []Protocol{discard},
Logger: testlog.Logger(t, log.LvlTrace),
}
srv := &Server{
Config: cfg,
newTransport: func(fd net.Conn, dialDest *ecdsa.PublicKey) transport { return test.tt },
log: cfg.Logger,
}
if !test.dontstart {
if err := srv.Start(); err != nil {
t.Fatalf("couldn't start server: %v", err)
}
defer srv.Stop()
}
p1, _ := net.Pipe()
srv.SetupConn(p1, test.flags, test.dialDest)
if !errors.Is(test.tt.closeErr, test.wantCloseErr) {
t.Errorf("test %d: close error mismatch: got %q, want %q", i, test.tt.closeErr, test.wantCloseErr)
}
if test.tt.calls != test.wantCalls {
t.Errorf("test %d: calls mismatch: got %q, want %q", i, test.tt.calls, test.wantCalls)
}
})
}
}
type setupTransport struct {
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
pubkey *ecdsa.PublicKey
encHandshakeErr error
phs protoHandshake
protoHandshakeErr error
calls string
closeErr error
}
func (c *setupTransport) doEncHandshake(prv *ecdsa.PrivateKey) (*ecdsa.PublicKey, error) {
c.calls += "doEncHandshake,"
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
return c.pubkey, c.encHandshakeErr
}
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
func (c *setupTransport) doProtoHandshake(our *protoHandshake) (*protoHandshake, error) {
c.calls += "doProtoHandshake,"
if c.protoHandshakeErr != nil {
return nil, c.protoHandshakeErr
}
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
return &c.phs, nil
}
func (c *setupTransport) close(err error) {
c.calls += "close,"
c.closeErr = err
}
// setupConn shouldn't write to/read from the connection.
func (c *setupTransport) WriteMsg(Msg) error {
panic("WriteMsg called on setupTransport")
}
func (c *setupTransport) ReadMsg() (Msg, error) {
panic("ReadMsg called on setupTransport")
}
func newkey() *ecdsa.PrivateKey {
key, err := crypto.GenerateKey()
if err != nil {
panic("couldn't generate key: " + err.Error())
}
return key
}
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-25 01:59:00 +03:00
func randomID() (id enode.ID) {
for i := range id {
id[i] = byte(rand.Intn(255))
}
return id
}
// This test checks that inbound connections are throttled by IP.
func TestServerInboundThrottle(t *testing.T) {
const timeout = 5 * time.Second
newTransportCalled := make(chan struct{})
srv := &Server{
Config: Config{
PrivateKey: newkey(),
ListenAddr: "127.0.0.1:0",
MaxPeers: 10,
NoDial: true,
NoDiscovery: true,
Protocols: []Protocol{discard},
Logger: testlog.Logger(t, log.LvlTrace),
},
newTransport: func(fd net.Conn, dialDest *ecdsa.PublicKey) transport {
newTransportCalled <- struct{}{}
return newRLPX(fd, dialDest)
},
listenFunc: func(network, laddr string) (net.Listener, error) {
fakeAddr := &net.TCPAddr{IP: net.IP{95, 33, 21, 2}, Port: 4444}
return listenFakeAddr(network, laddr, fakeAddr)
},
}
if err := srv.Start(); err != nil {
t.Fatal("can't start: ", err)
}
defer srv.Stop()
// Dial the test server.
conn, err := net.DialTimeout("tcp", srv.ListenAddr, timeout)
if err != nil {
t.Fatalf("could not dial: %v", err)
}
select {
case <-newTransportCalled:
// OK
case <-time.After(timeout):
t.Error("newTransport not called")
}
conn.Close()
// Dial again. This time the server should close the connection immediately.
connClosed := make(chan struct{}, 1)
conn, err = net.DialTimeout("tcp", srv.ListenAddr, timeout)
if err != nil {
t.Fatalf("could not dial: %v", err)
}
defer conn.Close()
go func() {
conn.SetDeadline(time.Now().Add(timeout))
buf := make([]byte, 10)
if n, err := conn.Read(buf); err != io.EOF || n != 0 {
t.Errorf("expected io.EOF and n == 0, got error %q and n == %d", err, n)
}
connClosed <- struct{}{}
}()
select {
case <-connClosed:
// OK
case <-newTransportCalled:
t.Error("newTransport called for second attempt")
case <-time.After(timeout):
t.Error("connection not closed within timeout")
}
}
func listenFakeAddr(network, laddr string, remoteAddr net.Addr) (net.Listener, error) {
l, err := net.Listen(network, laddr)
if err == nil {
l = &fakeAddrListener{l, remoteAddr}
}
return l, err
}
// fakeAddrListener is a listener that creates connections with a mocked remote address.
type fakeAddrListener struct {
net.Listener
remoteAddr net.Addr
}
type fakeAddrConn struct {
net.Conn
remoteAddr net.Addr
}
func (l *fakeAddrListener) Accept() (net.Conn, error) {
c, err := l.Listener.Accept()
if err != nil {
return nil, err
}
return &fakeAddrConn{c, l.remoteAddr}, nil
}
func (c *fakeAddrConn) RemoteAddr() net.Addr {
return c.remoteAddr
}
p2p: new dial scheduler (#20592) * p2p: new dial scheduler This change replaces the peer-to-peer dial scheduler with a new and improved implementation. The new code is better than the previous implementation in two key aspects: - The time between discovery of a node and dialing that node is significantly lower in the new version. The old dialState kept a buffer of nodes and launched a task to refill it whenever the buffer became empty. This worked well with the discovery interface we used to have, but doesn't really work with the new iterator-based discovery API. - Selection of static dial candidates (created by Server.AddPeer or through static-nodes.json) performs much better for large amounts of static peers. Connections to static nodes are now limited like dynanic dials and can no longer overstep MaxPeers or the dial ratio. * p2p/simulations/adapters: adapt to new NodeDialer interface * p2p: re-add check for self in checkDial * p2p: remove peersetCh * p2p: allow static dials when discovery is disabled * p2p: add test for dialScheduler.removeStatic * p2p: remove blank line * p2p: fix documentation of maxDialPeers * p2p: change "ok" to "added" in static node log * p2p: improve dialTask docs Also increase log level for "Can't resolve node" * p2p: ensure dial resolver is truly nil without discovery * p2p: add "looking for peers" log message * p2p: clean up Server.run comments * p2p: fix maxDialedConns for maxpeers < dialRatio Always allocate at least one dial slot unless dialing is disabled using NoDial or MaxPeers == 0. Most importantly, this fixes MaxPeers == 1 to dedicate the sole slot to dialing instead of listening. * p2p: fix RemovePeer to disconnect the peer again Also make RemovePeer synchronous and add a test. * p2p: remove "Connection set up" log message * p2p: clean up connection logging We previously logged outgoing connection failures up to three times. - in SetupConn() as "Setting up connection failed addr=..." - in setupConn() with an error-specific message and "id=... addr=..." - in dial() as "Dial error task=..." This commit ensures a single log message is emitted per failure and adds "id=... addr=... conn=..." everywhere (id= omitted when the ID isn't known yet). Also avoid printing a log message when a static dial fails but can't be resolved because discv4 is disabled. The light client hit this case all the time, increasing the message count to four lines per failed connection. * p2p: document that RemovePeer blocks
2020-02-13 13:10:03 +03:00
func syncAddPeer(srv *Server, node *enode.Node) bool {
var (
ch = make(chan *PeerEvent)
sub = srv.SubscribeEvents(ch)
timeout = time.After(2 * time.Second)
)
defer sub.Unsubscribe()
srv.AddPeer(node)
for {
select {
case ev := <-ch:
if ev.Type == PeerEventTypeAdd && ev.Peer == node.ID() {
return true
}
case <-timeout:
return false
}
}
}