2019-07-22 12:17:27 +03:00
|
|
|
// Copyright 2018 The go-ethereum Authors
|
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 00:59:00 +02:00
|
|
|
// This file is part of the go-ethereum library.
|
|
|
|
//
|
|
|
|
// The go-ethereum library is free software: you can redistribute it and/or modify
|
|
|
|
// 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,
|
|
|
|
// but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
// 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/>.
|
|
|
|
|
|
|
|
package discover
|
|
|
|
|
|
|
|
import (
|
2019-10-29 16:08:57 +01:00
|
|
|
"bytes"
|
2019-04-30 13:13:22 +02:00
|
|
|
"crypto/ecdsa"
|
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 00:59:00 +02:00
|
|
|
"encoding/hex"
|
2019-10-29 16:08:57 +01:00
|
|
|
"errors"
|
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 00:59:00 +02:00
|
|
|
"fmt"
|
|
|
|
"math/rand"
|
|
|
|
"net"
|
2024-03-25 07:50:18 +01:00
|
|
|
"slices"
|
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 00:59:00 +02:00
|
|
|
"sync"
|
p2p/discover: improved node revalidation (#29572)
Node discovery periodically revalidates the nodes in its table by sending PING, checking
if they are still alive. I recently noticed some issues with the implementation of this
process, which can cause strange results such as nodes dropping unexpectedly, certain
nodes not getting revalidated often enough, and bad results being returned to incoming
FINDNODE queries.
In this change, the revalidation process is improved with the following logic:
- We maintain two 'revalidation lists' containing the table nodes, named 'fast' and 'slow'.
- The process chooses random nodes from each list on a randomized interval, the interval being
faster for the 'fast' list, and performs revalidation for the chosen node.
- Whenever a node is newly inserted into the table, it goes into the 'fast' list.
Once validation passes, it transfers to the 'slow' list. If a request fails, or the
node changes endpoint, it transfers back into 'fast'.
- livenessChecks is incremented by one for successful checks. Unlike the old implementation,
we will not drop the node on the first failing check. We instead quickly decay the
livenessChecks give it another chance.
- Order of nodes in bucket doesn't matter anymore.
I am also adding a debug API endpoint to dump the node table content.
Co-authored-by: Martin HS <martin@swende.se>
2024-05-23 14:26:09 +02:00
|
|
|
"sync/atomic"
|
|
|
|
"time"
|
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 00:59:00 +02:00
|
|
|
|
2019-04-30 13:13:22 +02:00
|
|
|
"github.com/ethereum/go-ethereum/crypto"
|
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 00:59:00 +02:00
|
|
|
"github.com/ethereum/go-ethereum/p2p/enode"
|
|
|
|
"github.com/ethereum/go-ethereum/p2p/enr"
|
|
|
|
)
|
|
|
|
|
2018-10-12 11:47:24 +02:00
|
|
|
var nullNode *enode.Node
|
|
|
|
|
|
|
|
func init() {
|
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 00:59:00 +02:00
|
|
|
var r enr.Record
|
|
|
|
r.Set(enr.IP{0, 0, 0, 0})
|
2018-10-12 11:47:24 +02:00
|
|
|
nullNode = enode.SignNull(&r, enode.ID{})
|
|
|
|
}
|
|
|
|
|
p2p/discover: improved node revalidation (#29572)
Node discovery periodically revalidates the nodes in its table by sending PING, checking
if they are still alive. I recently noticed some issues with the implementation of this
process, which can cause strange results such as nodes dropping unexpectedly, certain
nodes not getting revalidated often enough, and bad results being returned to incoming
FINDNODE queries.
In this change, the revalidation process is improved with the following logic:
- We maintain two 'revalidation lists' containing the table nodes, named 'fast' and 'slow'.
- The process chooses random nodes from each list on a randomized interval, the interval being
faster for the 'fast' list, and performs revalidation for the chosen node.
- Whenever a node is newly inserted into the table, it goes into the 'fast' list.
Once validation passes, it transfers to the 'slow' list. If a request fails, or the
node changes endpoint, it transfers back into 'fast'.
- livenessChecks is incremented by one for successful checks. Unlike the old implementation,
we will not drop the node on the first failing check. We instead quickly decay the
livenessChecks give it another chance.
- Order of nodes in bucket doesn't matter anymore.
I am also adding a debug API endpoint to dump the node table content.
Co-authored-by: Martin HS <martin@swende.se>
2024-05-23 14:26:09 +02:00
|
|
|
func newTestTable(t transport, cfg Config) (*Table, *enode.DB) {
|
2024-05-28 18:13:03 +02:00
|
|
|
tab, db := newInactiveTestTable(t, cfg)
|
|
|
|
go tab.loop()
|
|
|
|
return tab, db
|
|
|
|
}
|
|
|
|
|
|
|
|
// newInactiveTestTable creates a Table without running the main loop.
|
|
|
|
func newInactiveTestTable(t transport, cfg Config) (*Table, *enode.DB) {
|
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 00:59:00 +02:00
|
|
|
db, _ := enode.OpenDB("")
|
2023-05-31 13:37:10 +02:00
|
|
|
tab, _ := newTable(t, db, cfg)
|
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 00:59:00 +02:00
|
|
|
return tab, db
|
|
|
|
}
|
|
|
|
|
|
|
|
// nodeAtDistance creates a node for which enode.LogDist(base, n.id) == ld.
|
2024-05-29 15:02:26 +02:00
|
|
|
func nodeAtDistance(base enode.ID, ld int, ip net.IP) *enode.Node {
|
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 00:59:00 +02:00
|
|
|
var r enr.Record
|
|
|
|
r.Set(enr.IP(ip))
|
2023-04-28 11:03:43 +02:00
|
|
|
r.Set(enr.UDP(30303))
|
2024-05-29 15:02:26 +02:00
|
|
|
return enode.SignNull(&r, idAtDistance(base, ld))
|
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 00:59:00 +02:00
|
|
|
}
|
|
|
|
|
2020-04-08 09:57:23 +02:00
|
|
|
// nodesAtDistance creates n nodes for which enode.LogDist(base, node.ID()) == ld.
|
|
|
|
func nodesAtDistance(base enode.ID, ld int, n int) []*enode.Node {
|
|
|
|
results := make([]*enode.Node, n)
|
|
|
|
for i := range results {
|
2024-05-29 15:02:26 +02:00
|
|
|
results[i] = nodeAtDistance(base, ld, intIP(i))
|
2020-04-08 09:57:23 +02:00
|
|
|
}
|
|
|
|
return results
|
|
|
|
}
|
|
|
|
|
|
|
|
func nodesToRecords(nodes []*enode.Node) []*enr.Record {
|
|
|
|
records := make([]*enr.Record, len(nodes))
|
|
|
|
for i := range nodes {
|
|
|
|
records[i] = nodes[i].Record()
|
|
|
|
}
|
|
|
|
return records
|
|
|
|
}
|
|
|
|
|
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 00:59:00 +02:00
|
|
|
// idAtDistance returns a random hash such that enode.LogDist(a, b) == n
|
|
|
|
func idAtDistance(a enode.ID, n int) (b enode.ID) {
|
|
|
|
if n == 0 {
|
|
|
|
return a
|
|
|
|
}
|
|
|
|
// flip bit at position n, fill the rest with random bits
|
|
|
|
b = a
|
|
|
|
pos := len(a) - n/8 - 1
|
|
|
|
bit := byte(0x01) << (byte(n%8) - 1)
|
|
|
|
if bit == 0 {
|
|
|
|
pos++
|
|
|
|
bit = 0x80
|
|
|
|
}
|
|
|
|
b[pos] = a[pos]&^bit | ^a[pos]&bit // TODO: randomize end bits
|
|
|
|
for i := pos + 1; i < len(a); i++ {
|
|
|
|
b[i] = byte(rand.Intn(255))
|
|
|
|
}
|
|
|
|
return b
|
|
|
|
}
|
|
|
|
|
|
|
|
func intIP(i int) net.IP {
|
|
|
|
return net.IP{byte(i), 0, 2, byte(i)}
|
|
|
|
}
|
|
|
|
|
|
|
|
// fillBucket inserts nodes into the given bucket until it is full.
|
2024-05-29 15:02:26 +02:00
|
|
|
func fillBucket(tab *Table, id enode.ID) (last *tableNode) {
|
p2p/discover: improved node revalidation (#29572)
Node discovery periodically revalidates the nodes in its table by sending PING, checking
if they are still alive. I recently noticed some issues with the implementation of this
process, which can cause strange results such as nodes dropping unexpectedly, certain
nodes not getting revalidated often enough, and bad results being returned to incoming
FINDNODE queries.
In this change, the revalidation process is improved with the following logic:
- We maintain two 'revalidation lists' containing the table nodes, named 'fast' and 'slow'.
- The process chooses random nodes from each list on a randomized interval, the interval being
faster for the 'fast' list, and performs revalidation for the chosen node.
- Whenever a node is newly inserted into the table, it goes into the 'fast' list.
Once validation passes, it transfers to the 'slow' list. If a request fails, or the
node changes endpoint, it transfers back into 'fast'.
- livenessChecks is incremented by one for successful checks. Unlike the old implementation,
we will not drop the node on the first failing check. We instead quickly decay the
livenessChecks give it another chance.
- Order of nodes in bucket doesn't matter anymore.
I am also adding a debug API endpoint to dump the node table content.
Co-authored-by: Martin HS <martin@swende.se>
2024-05-23 14:26:09 +02:00
|
|
|
ld := enode.LogDist(tab.self().ID(), id)
|
|
|
|
b := tab.bucket(id)
|
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 00:59:00 +02:00
|
|
|
for len(b.entries) < bucketSize {
|
p2p/discover: improved node revalidation (#29572)
Node discovery periodically revalidates the nodes in its table by sending PING, checking
if they are still alive. I recently noticed some issues with the implementation of this
process, which can cause strange results such as nodes dropping unexpectedly, certain
nodes not getting revalidated often enough, and bad results being returned to incoming
FINDNODE queries.
In this change, the revalidation process is improved with the following logic:
- We maintain two 'revalidation lists' containing the table nodes, named 'fast' and 'slow'.
- The process chooses random nodes from each list on a randomized interval, the interval being
faster for the 'fast' list, and performs revalidation for the chosen node.
- Whenever a node is newly inserted into the table, it goes into the 'fast' list.
Once validation passes, it transfers to the 'slow' list. If a request fails, or the
node changes endpoint, it transfers back into 'fast'.
- livenessChecks is incremented by one for successful checks. Unlike the old implementation,
we will not drop the node on the first failing check. We instead quickly decay the
livenessChecks give it another chance.
- Order of nodes in bucket doesn't matter anymore.
I am also adding a debug API endpoint to dump the node table content.
Co-authored-by: Martin HS <martin@swende.se>
2024-05-23 14:26:09 +02:00
|
|
|
node := nodeAtDistance(tab.self().ID(), ld, intIP(ld))
|
2024-05-29 15:02:26 +02:00
|
|
|
if !tab.addFoundNode(node, false) {
|
p2p/discover: improved node revalidation (#29572)
Node discovery periodically revalidates the nodes in its table by sending PING, checking
if they are still alive. I recently noticed some issues with the implementation of this
process, which can cause strange results such as nodes dropping unexpectedly, certain
nodes not getting revalidated often enough, and bad results being returned to incoming
FINDNODE queries.
In this change, the revalidation process is improved with the following logic:
- We maintain two 'revalidation lists' containing the table nodes, named 'fast' and 'slow'.
- The process chooses random nodes from each list on a randomized interval, the interval being
faster for the 'fast' list, and performs revalidation for the chosen node.
- Whenever a node is newly inserted into the table, it goes into the 'fast' list.
Once validation passes, it transfers to the 'slow' list. If a request fails, or the
node changes endpoint, it transfers back into 'fast'.
- livenessChecks is incremented by one for successful checks. Unlike the old implementation,
we will not drop the node on the first failing check. We instead quickly decay the
livenessChecks give it another chance.
- Order of nodes in bucket doesn't matter anymore.
I am also adding a debug API endpoint to dump the node table content.
Co-authored-by: Martin HS <martin@swende.se>
2024-05-23 14:26:09 +02:00
|
|
|
panic("node not added")
|
|
|
|
}
|
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 00:59:00 +02:00
|
|
|
}
|
|
|
|
return b.entries[bucketSize-1]
|
|
|
|
}
|
|
|
|
|
2019-01-29 17:39:20 +01:00
|
|
|
// fillTable adds nodes the table to the end of their corresponding bucket
|
|
|
|
// if the bucket is not full. The caller must not hold tab.mutex.
|
2024-05-29 15:02:26 +02:00
|
|
|
func fillTable(tab *Table, nodes []*enode.Node, setLive bool) {
|
2019-01-29 17:39:20 +01:00
|
|
|
for _, n := range nodes {
|
2024-05-29 15:02:26 +02:00
|
|
|
tab.addFoundNode(n, setLive)
|
2019-01-29 17:39:20 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
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 00:59:00 +02:00
|
|
|
type pingRecorder struct {
|
p2p/discover: improved node revalidation (#29572)
Node discovery periodically revalidates the nodes in its table by sending PING, checking
if they are still alive. I recently noticed some issues with the implementation of this
process, which can cause strange results such as nodes dropping unexpectedly, certain
nodes not getting revalidated often enough, and bad results being returned to incoming
FINDNODE queries.
In this change, the revalidation process is improved with the following logic:
- We maintain two 'revalidation lists' containing the table nodes, named 'fast' and 'slow'.
- The process chooses random nodes from each list on a randomized interval, the interval being
faster for the 'fast' list, and performs revalidation for the chosen node.
- Whenever a node is newly inserted into the table, it goes into the 'fast' list.
Once validation passes, it transfers to the 'slow' list. If a request fails, or the
node changes endpoint, it transfers back into 'fast'.
- livenessChecks is incremented by one for successful checks. Unlike the old implementation,
we will not drop the node on the first failing check. We instead quickly decay the
livenessChecks give it another chance.
- Order of nodes in bucket doesn't matter anymore.
I am also adding a debug API endpoint to dump the node table content.
Co-authored-by: Martin HS <martin@swende.se>
2024-05-23 14:26:09 +02:00
|
|
|
mu sync.Mutex
|
|
|
|
cond *sync.Cond
|
|
|
|
dead map[enode.ID]bool
|
|
|
|
records map[enode.ID]*enode.Node
|
|
|
|
pinged []*enode.Node
|
|
|
|
n *enode.Node
|
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 00:59:00 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
func newPingRecorder() *pingRecorder {
|
2018-10-12 11:47:24 +02:00
|
|
|
var r enr.Record
|
|
|
|
r.Set(enr.IP{0, 0, 0, 0})
|
|
|
|
n := enode.SignNull(&r, enode.ID{})
|
|
|
|
|
p2p/discover: improved node revalidation (#29572)
Node discovery periodically revalidates the nodes in its table by sending PING, checking
if they are still alive. I recently noticed some issues with the implementation of this
process, which can cause strange results such as nodes dropping unexpectedly, certain
nodes not getting revalidated often enough, and bad results being returned to incoming
FINDNODE queries.
In this change, the revalidation process is improved with the following logic:
- We maintain two 'revalidation lists' containing the table nodes, named 'fast' and 'slow'.
- The process chooses random nodes from each list on a randomized interval, the interval being
faster for the 'fast' list, and performs revalidation for the chosen node.
- Whenever a node is newly inserted into the table, it goes into the 'fast' list.
Once validation passes, it transfers to the 'slow' list. If a request fails, or the
node changes endpoint, it transfers back into 'fast'.
- livenessChecks is incremented by one for successful checks. Unlike the old implementation,
we will not drop the node on the first failing check. We instead quickly decay the
livenessChecks give it another chance.
- Order of nodes in bucket doesn't matter anymore.
I am also adding a debug API endpoint to dump the node table content.
Co-authored-by: Martin HS <martin@swende.se>
2024-05-23 14:26:09 +02:00
|
|
|
t := &pingRecorder{
|
2019-05-15 06:47:45 +02:00
|
|
|
dead: make(map[enode.ID]bool),
|
|
|
|
records: make(map[enode.ID]*enode.Node),
|
|
|
|
n: n,
|
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 00:59:00 +02:00
|
|
|
}
|
p2p/discover: improved node revalidation (#29572)
Node discovery periodically revalidates the nodes in its table by sending PING, checking
if they are still alive. I recently noticed some issues with the implementation of this
process, which can cause strange results such as nodes dropping unexpectedly, certain
nodes not getting revalidated often enough, and bad results being returned to incoming
FINDNODE queries.
In this change, the revalidation process is improved with the following logic:
- We maintain two 'revalidation lists' containing the table nodes, named 'fast' and 'slow'.
- The process chooses random nodes from each list on a randomized interval, the interval being
faster for the 'fast' list, and performs revalidation for the chosen node.
- Whenever a node is newly inserted into the table, it goes into the 'fast' list.
Once validation passes, it transfers to the 'slow' list. If a request fails, or the
node changes endpoint, it transfers back into 'fast'.
- livenessChecks is incremented by one for successful checks. Unlike the old implementation,
we will not drop the node on the first failing check. We instead quickly decay the
livenessChecks give it another chance.
- Order of nodes in bucket doesn't matter anymore.
I am also adding a debug API endpoint to dump the node table content.
Co-authored-by: Martin HS <martin@swende.se>
2024-05-23 14:26:09 +02:00
|
|
|
t.cond = sync.NewCond(&t.mu)
|
|
|
|
return 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 00:59:00 +02:00
|
|
|
}
|
|
|
|
|
2022-07-10 17:15:54 +09:00
|
|
|
// updateRecord updates a node record. Future calls to ping and
|
|
|
|
// RequestENR will return this record.
|
2019-05-15 06:47:45 +02:00
|
|
|
func (t *pingRecorder) updateRecord(n *enode.Node) {
|
|
|
|
t.mu.Lock()
|
|
|
|
defer t.mu.Unlock()
|
|
|
|
t.records[n.ID()] = n
|
2018-10-12 11:47:24 +02:00
|
|
|
}
|
|
|
|
|
2019-05-15 06:47:45 +02:00
|
|
|
// Stubs to satisfy the transport interface.
|
|
|
|
func (t *pingRecorder) Self() *enode.Node { return nullNode }
|
|
|
|
func (t *pingRecorder) lookupSelf() []*enode.Node { return nil }
|
|
|
|
func (t *pingRecorder) lookupRandom() []*enode.Node { return nil }
|
|
|
|
|
p2p/discover: improved node revalidation (#29572)
Node discovery periodically revalidates the nodes in its table by sending PING, checking
if they are still alive. I recently noticed some issues with the implementation of this
process, which can cause strange results such as nodes dropping unexpectedly, certain
nodes not getting revalidated often enough, and bad results being returned to incoming
FINDNODE queries.
In this change, the revalidation process is improved with the following logic:
- We maintain two 'revalidation lists' containing the table nodes, named 'fast' and 'slow'.
- The process chooses random nodes from each list on a randomized interval, the interval being
faster for the 'fast' list, and performs revalidation for the chosen node.
- Whenever a node is newly inserted into the table, it goes into the 'fast' list.
Once validation passes, it transfers to the 'slow' list. If a request fails, or the
node changes endpoint, it transfers back into 'fast'.
- livenessChecks is incremented by one for successful checks. Unlike the old implementation,
we will not drop the node on the first failing check. We instead quickly decay the
livenessChecks give it another chance.
- Order of nodes in bucket doesn't matter anymore.
I am also adding a debug API endpoint to dump the node table content.
Co-authored-by: Martin HS <martin@swende.se>
2024-05-23 14:26:09 +02:00
|
|
|
func (t *pingRecorder) waitPing(timeout time.Duration) *enode.Node {
|
|
|
|
t.mu.Lock()
|
|
|
|
defer t.mu.Unlock()
|
|
|
|
|
|
|
|
// Wake up the loop on timeout.
|
|
|
|
var timedout atomic.Bool
|
|
|
|
timer := time.AfterFunc(timeout, func() {
|
|
|
|
timedout.Store(true)
|
|
|
|
t.cond.Broadcast()
|
|
|
|
})
|
|
|
|
defer timer.Stop()
|
|
|
|
|
|
|
|
// Wait for a ping.
|
|
|
|
for {
|
|
|
|
if timedout.Load() {
|
|
|
|
return nil
|
|
|
|
}
|
|
|
|
if len(t.pinged) > 0 {
|
|
|
|
n := t.pinged[0]
|
|
|
|
t.pinged = append(t.pinged[:0], t.pinged[1:]...)
|
|
|
|
return n
|
|
|
|
}
|
|
|
|
t.cond.Wait()
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2019-05-15 06:47:45 +02:00
|
|
|
// ping simulates a ping request.
|
|
|
|
func (t *pingRecorder) ping(n *enode.Node) (seq uint64, err 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 00:59:00 +02:00
|
|
|
t.mu.Lock()
|
|
|
|
defer t.mu.Unlock()
|
|
|
|
|
p2p/discover: improved node revalidation (#29572)
Node discovery periodically revalidates the nodes in its table by sending PING, checking
if they are still alive. I recently noticed some issues with the implementation of this
process, which can cause strange results such as nodes dropping unexpectedly, certain
nodes not getting revalidated often enough, and bad results being returned to incoming
FINDNODE queries.
In this change, the revalidation process is improved with the following logic:
- We maintain two 'revalidation lists' containing the table nodes, named 'fast' and 'slow'.
- The process chooses random nodes from each list on a randomized interval, the interval being
faster for the 'fast' list, and performs revalidation for the chosen node.
- Whenever a node is newly inserted into the table, it goes into the 'fast' list.
Once validation passes, it transfers to the 'slow' list. If a request fails, or the
node changes endpoint, it transfers back into 'fast'.
- livenessChecks is incremented by one for successful checks. Unlike the old implementation,
we will not drop the node on the first failing check. We instead quickly decay the
livenessChecks give it another chance.
- Order of nodes in bucket doesn't matter anymore.
I am also adding a debug API endpoint to dump the node table content.
Co-authored-by: Martin HS <martin@swende.se>
2024-05-23 14:26:09 +02:00
|
|
|
t.pinged = append(t.pinged, n)
|
|
|
|
t.cond.Broadcast()
|
|
|
|
|
2019-04-30 13:13:22 +02:00
|
|
|
if t.dead[n.ID()] {
|
2019-05-15 06:47:45 +02:00
|
|
|
return 0, errTimeout
|
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 00:59:00 +02:00
|
|
|
}
|
2019-05-15 06:47:45 +02:00
|
|
|
if t.records[n.ID()] != nil {
|
|
|
|
seq = t.records[n.ID()].Seq()
|
|
|
|
}
|
|
|
|
return seq, 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 00:59:00 +02:00
|
|
|
}
|
|
|
|
|
2022-07-10 17:15:54 +09:00
|
|
|
// RequestENR simulates an ENR request.
|
2019-06-07 15:29:16 +02:00
|
|
|
func (t *pingRecorder) RequestENR(n *enode.Node) (*enode.Node, error) {
|
2019-05-15 06:47:45 +02:00
|
|
|
t.mu.Lock()
|
|
|
|
defer t.mu.Unlock()
|
2019-04-30 13:13:22 +02:00
|
|
|
|
2019-05-15 06:47:45 +02:00
|
|
|
if t.dead[n.ID()] || t.records[n.ID()] == nil {
|
|
|
|
return nil, errTimeout
|
|
|
|
}
|
|
|
|
return t.records[n.ID()], nil
|
2019-04-30 13:13:22 +02:00
|
|
|
}
|
|
|
|
|
2024-05-29 15:02:26 +02:00
|
|
|
func hasDuplicates(slice []*enode.Node) bool {
|
2023-05-10 02:52:26 -06:00
|
|
|
seen := make(map[enode.ID]bool, len(slice))
|
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 00:59:00 +02:00
|
|
|
for i, e := range slice {
|
|
|
|
if e == nil {
|
|
|
|
panic(fmt.Sprintf("nil *Node at %d", i))
|
|
|
|
}
|
|
|
|
if seen[e.ID()] {
|
|
|
|
return true
|
|
|
|
}
|
|
|
|
seen[e.ID()] = true
|
|
|
|
}
|
|
|
|
return false
|
|
|
|
}
|
|
|
|
|
2020-10-14 12:28:17 +02:00
|
|
|
// checkNodesEqual checks whether the two given node lists contain the same nodes.
|
2019-10-29 16:08:57 +01:00
|
|
|
func checkNodesEqual(got, want []*enode.Node) error {
|
2020-04-08 09:57:23 +02:00
|
|
|
if len(got) == len(want) {
|
|
|
|
for i := range got {
|
|
|
|
if !nodeEqual(got[i], want[i]) {
|
|
|
|
goto NotEqual
|
|
|
|
}
|
|
|
|
}
|
2019-10-29 16:08:57 +01:00
|
|
|
}
|
2020-10-14 12:28:17 +02:00
|
|
|
return nil
|
2020-04-08 09:57:23 +02:00
|
|
|
|
|
|
|
NotEqual:
|
2019-10-29 16:08:57 +01:00
|
|
|
output := new(bytes.Buffer)
|
|
|
|
fmt.Fprintf(output, "got %d nodes:\n", len(got))
|
|
|
|
for _, n := range got {
|
|
|
|
fmt.Fprintf(output, " %v %v\n", n.ID(), n)
|
|
|
|
}
|
|
|
|
fmt.Fprintf(output, "want %d:\n", len(want))
|
|
|
|
for _, n := range want {
|
|
|
|
fmt.Fprintf(output, " %v %v\n", n.ID(), n)
|
|
|
|
}
|
|
|
|
return errors.New(output.String())
|
|
|
|
}
|
|
|
|
|
2020-04-08 09:57:23 +02:00
|
|
|
func nodeEqual(n1 *enode.Node, n2 *enode.Node) bool {
|
|
|
|
return n1.ID() == n2.ID() && n1.IP().Equal(n2.IP())
|
|
|
|
}
|
|
|
|
|
2024-05-29 15:02:26 +02:00
|
|
|
func sortByID[N nodeType](nodes []N) {
|
|
|
|
slices.SortFunc(nodes, func(a, b N) int {
|
2023-08-12 01:04:12 +03:00
|
|
|
return bytes.Compare(a.ID().Bytes(), b.ID().Bytes())
|
2019-10-29 16:08:57 +01:00
|
|
|
})
|
|
|
|
}
|
|
|
|
|
2024-05-29 15:02:26 +02:00
|
|
|
func sortedByDistanceTo(distbase enode.ID, slice []*enode.Node) bool {
|
|
|
|
return slices.IsSortedFunc(slice, func(a, b *enode.Node) int {
|
2023-08-12 01:04:12 +03:00
|
|
|
return enode.DistCmp(distbase, a.ID(), b.ID())
|
2019-04-30 13:13:22 +02:00
|
|
|
})
|
|
|
|
}
|
|
|
|
|
2020-10-14 12:28:17 +02:00
|
|
|
// hexEncPrivkey decodes h as a private key.
|
2019-04-30 13:13:22 +02:00
|
|
|
func hexEncPrivkey(h string) *ecdsa.PrivateKey {
|
|
|
|
b, err := hex.DecodeString(h)
|
|
|
|
if err != nil {
|
|
|
|
panic(err)
|
|
|
|
}
|
|
|
|
key, err := crypto.ToECDSA(b)
|
|
|
|
if err != nil {
|
|
|
|
panic(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 00:59:00 +02:00
|
|
|
}
|
2019-04-30 13:13:22 +02:00
|
|
|
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 00:59:00 +02:00
|
|
|
}
|
|
|
|
|
2020-10-14 12:28:17 +02:00
|
|
|
// hexEncPubkey decodes h as a public 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 00:59:00 +02:00
|
|
|
func hexEncPubkey(h string) (ret encPubkey) {
|
|
|
|
b, err := hex.DecodeString(h)
|
|
|
|
if err != nil {
|
|
|
|
panic(err)
|
|
|
|
}
|
|
|
|
if len(b) != len(ret) {
|
|
|
|
panic("invalid length")
|
|
|
|
}
|
|
|
|
copy(ret[:], b)
|
|
|
|
return ret
|
|
|
|
}
|
p2p/discover: improved node revalidation (#29572)
Node discovery periodically revalidates the nodes in its table by sending PING, checking
if they are still alive. I recently noticed some issues with the implementation of this
process, which can cause strange results such as nodes dropping unexpectedly, certain
nodes not getting revalidated often enough, and bad results being returned to incoming
FINDNODE queries.
In this change, the revalidation process is improved with the following logic:
- We maintain two 'revalidation lists' containing the table nodes, named 'fast' and 'slow'.
- The process chooses random nodes from each list on a randomized interval, the interval being
faster for the 'fast' list, and performs revalidation for the chosen node.
- Whenever a node is newly inserted into the table, it goes into the 'fast' list.
Once validation passes, it transfers to the 'slow' list. If a request fails, or the
node changes endpoint, it transfers back into 'fast'.
- livenessChecks is incremented by one for successful checks. Unlike the old implementation,
we will not drop the node on the first failing check. We instead quickly decay the
livenessChecks give it another chance.
- Order of nodes in bucket doesn't matter anymore.
I am also adding a debug API endpoint to dump the node table content.
Co-authored-by: Martin HS <martin@swende.se>
2024-05-23 14:26:09 +02:00
|
|
|
|
|
|
|
type nodeEventRecorder struct {
|
|
|
|
evc chan recordedNodeEvent
|
|
|
|
}
|
|
|
|
|
|
|
|
type recordedNodeEvent struct {
|
2024-05-29 15:02:26 +02:00
|
|
|
node *tableNode
|
p2p/discover: improved node revalidation (#29572)
Node discovery periodically revalidates the nodes in its table by sending PING, checking
if they are still alive. I recently noticed some issues with the implementation of this
process, which can cause strange results such as nodes dropping unexpectedly, certain
nodes not getting revalidated often enough, and bad results being returned to incoming
FINDNODE queries.
In this change, the revalidation process is improved with the following logic:
- We maintain two 'revalidation lists' containing the table nodes, named 'fast' and 'slow'.
- The process chooses random nodes from each list on a randomized interval, the interval being
faster for the 'fast' list, and performs revalidation for the chosen node.
- Whenever a node is newly inserted into the table, it goes into the 'fast' list.
Once validation passes, it transfers to the 'slow' list. If a request fails, or the
node changes endpoint, it transfers back into 'fast'.
- livenessChecks is incremented by one for successful checks. Unlike the old implementation,
we will not drop the node on the first failing check. We instead quickly decay the
livenessChecks give it another chance.
- Order of nodes in bucket doesn't matter anymore.
I am also adding a debug API endpoint to dump the node table content.
Co-authored-by: Martin HS <martin@swende.se>
2024-05-23 14:26:09 +02:00
|
|
|
added bool
|
|
|
|
}
|
|
|
|
|
|
|
|
func newNodeEventRecorder(buffer int) *nodeEventRecorder {
|
|
|
|
return &nodeEventRecorder{
|
|
|
|
evc: make(chan recordedNodeEvent, buffer),
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2024-05-29 15:02:26 +02:00
|
|
|
func (set *nodeEventRecorder) nodeAdded(b *bucket, n *tableNode) {
|
p2p/discover: improved node revalidation (#29572)
Node discovery periodically revalidates the nodes in its table by sending PING, checking
if they are still alive. I recently noticed some issues with the implementation of this
process, which can cause strange results such as nodes dropping unexpectedly, certain
nodes not getting revalidated often enough, and bad results being returned to incoming
FINDNODE queries.
In this change, the revalidation process is improved with the following logic:
- We maintain two 'revalidation lists' containing the table nodes, named 'fast' and 'slow'.
- The process chooses random nodes from each list on a randomized interval, the interval being
faster for the 'fast' list, and performs revalidation for the chosen node.
- Whenever a node is newly inserted into the table, it goes into the 'fast' list.
Once validation passes, it transfers to the 'slow' list. If a request fails, or the
node changes endpoint, it transfers back into 'fast'.
- livenessChecks is incremented by one for successful checks. Unlike the old implementation,
we will not drop the node on the first failing check. We instead quickly decay the
livenessChecks give it another chance.
- Order of nodes in bucket doesn't matter anymore.
I am also adding a debug API endpoint to dump the node table content.
Co-authored-by: Martin HS <martin@swende.se>
2024-05-23 14:26:09 +02:00
|
|
|
select {
|
|
|
|
case set.evc <- recordedNodeEvent{n, true}:
|
|
|
|
default:
|
|
|
|
panic("no space in event buffer")
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2024-05-29 15:02:26 +02:00
|
|
|
func (set *nodeEventRecorder) nodeRemoved(b *bucket, n *tableNode) {
|
p2p/discover: improved node revalidation (#29572)
Node discovery periodically revalidates the nodes in its table by sending PING, checking
if they are still alive. I recently noticed some issues with the implementation of this
process, which can cause strange results such as nodes dropping unexpectedly, certain
nodes not getting revalidated often enough, and bad results being returned to incoming
FINDNODE queries.
In this change, the revalidation process is improved with the following logic:
- We maintain two 'revalidation lists' containing the table nodes, named 'fast' and 'slow'.
- The process chooses random nodes from each list on a randomized interval, the interval being
faster for the 'fast' list, and performs revalidation for the chosen node.
- Whenever a node is newly inserted into the table, it goes into the 'fast' list.
Once validation passes, it transfers to the 'slow' list. If a request fails, or the
node changes endpoint, it transfers back into 'fast'.
- livenessChecks is incremented by one for successful checks. Unlike the old implementation,
we will not drop the node on the first failing check. We instead quickly decay the
livenessChecks give it another chance.
- Order of nodes in bucket doesn't matter anymore.
I am also adding a debug API endpoint to dump the node table content.
Co-authored-by: Martin HS <martin@swende.se>
2024-05-23 14:26:09 +02:00
|
|
|
select {
|
|
|
|
case set.evc <- recordedNodeEvent{n, false}:
|
|
|
|
default:
|
|
|
|
panic("no space in event buffer")
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
func (set *nodeEventRecorder) waitNodePresent(id enode.ID, timeout time.Duration) bool {
|
|
|
|
return set.waitNodeEvent(id, timeout, true)
|
|
|
|
}
|
|
|
|
|
|
|
|
func (set *nodeEventRecorder) waitNodeAbsent(id enode.ID, timeout time.Duration) bool {
|
|
|
|
return set.waitNodeEvent(id, timeout, false)
|
|
|
|
}
|
|
|
|
|
|
|
|
func (set *nodeEventRecorder) waitNodeEvent(id enode.ID, timeout time.Duration, added bool) bool {
|
|
|
|
timer := time.NewTimer(timeout)
|
|
|
|
defer timer.Stop()
|
|
|
|
for {
|
|
|
|
select {
|
|
|
|
case ev := <-set.evc:
|
|
|
|
if ev.node.ID() == id && ev.added == added {
|
|
|
|
return true
|
|
|
|
}
|
|
|
|
case <-timer.C:
|
|
|
|
return false
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|