6c918a1806
The commit ensures that for every channel, there will always be two entries in the edges bucket. If the policy from one or both ends of the channel is unknown, it is marked as such. This allows efficient lookup of incoming edges. This is required for backwards payment path finding.
3068 lines
90 KiB
Go
3068 lines
90 KiB
Go
package channeldb
|
|
|
|
import (
|
|
"bytes"
|
|
"crypto/sha256"
|
|
"encoding/binary"
|
|
"fmt"
|
|
"image/color"
|
|
"io"
|
|
"math"
|
|
"net"
|
|
"time"
|
|
|
|
"github.com/btcsuite/btcd/btcec"
|
|
"github.com/btcsuite/btcd/chaincfg/chainhash"
|
|
"github.com/btcsuite/btcd/txscript"
|
|
"github.com/btcsuite/btcd/wire"
|
|
"github.com/btcsuite/btcutil"
|
|
"github.com/coreos/bbolt"
|
|
"github.com/lightningnetwork/lnd/lnwire"
|
|
)
|
|
|
|
var (
|
|
// nodeBucket is a bucket which houses all the vertices or nodes within
|
|
// the channel graph. This bucket has a single-sub bucket which adds an
|
|
// additional index from pubkey -> alias. Within the top-level of this
|
|
// bucket, the key space maps a node's compressed public key to the
|
|
// serialized information for that node. Additionally, there's a
|
|
// special key "source" which stores the pubkey of the source node. The
|
|
// source node is used as the starting point for all graph/queries and
|
|
// traversals. The graph is formed as a star-graph with the source node
|
|
// at the center.
|
|
//
|
|
// maps: pubKey -> nodeInfo
|
|
// maps: source -> selfPubKey
|
|
nodeBucket = []byte("graph-node")
|
|
|
|
// nodeUpdateIndexBucket is a sub-bucket of the nodeBucket. This bucket
|
|
// will be used to quickly look up the "freshness" of a node's last
|
|
// update to the network. The bucket only contains keys, and no values,
|
|
// it's mapping:
|
|
//
|
|
// maps: updateTime || nodeID -> nil
|
|
nodeUpdateIndexBucket = []byte("graph-node-update-index")
|
|
|
|
// sourceKey is a special key that resides within the nodeBucket. The
|
|
// sourceKey maps a key to the public key of the "self node".
|
|
sourceKey = []byte("source")
|
|
|
|
// aliasIndexBucket is a sub-bucket that's nested within the main
|
|
// nodeBucket. This bucket maps the public key of a node to its
|
|
// current alias. This bucket is provided as it can be used within a
|
|
// future UI layer to add an additional degree of confirmation.
|
|
aliasIndexBucket = []byte("alias")
|
|
|
|
// edgeBucket is a bucket which houses all of the edge or channel
|
|
// information within the channel graph. This bucket essentially acts
|
|
// as an adjacency list, which in conjunction with a range scan, can be
|
|
// used to iterate over all the incoming and outgoing edges for a
|
|
// particular node. Key in the bucket use a prefix scheme which leads
|
|
// with the node's public key and sends with the compact edge ID.
|
|
// For each chanID, there will be two entries within the bucket, as the
|
|
// graph is directed: nodes may have different policies w.r.t to fees
|
|
// for their respective directions.
|
|
//
|
|
// maps: pubKey || chanID -> channel edge policy for node
|
|
edgeBucket = []byte("graph-edge")
|
|
|
|
// unknownPolicy is represented as an empty slice. It is
|
|
// used as the value in edgeBucket for unknown channel edge policies.
|
|
// Unknown policies are still stored in the database to enable efficient
|
|
// lookup of incoming channel edges.
|
|
unknownPolicy = []byte{}
|
|
|
|
// chanStart is an array of all zero bytes which is used to perform
|
|
// range scans within the edgeBucket to obtain all of the outgoing
|
|
// edges for a particular node.
|
|
chanStart [8]byte
|
|
|
|
// edgeIndexBucket is an index which can be used to iterate all edges
|
|
// in the bucket, grouping them according to their in/out nodes.
|
|
// Additionally, the items in this bucket also contain the complete
|
|
// edge information for a channel. The edge information includes the
|
|
// capacity of the channel, the nodes that made the channel, etc. This
|
|
// bucket resides within the edgeBucket above. Creation of an edge
|
|
// proceeds in two phases: first the edge is added to the edge index,
|
|
// afterwards the edgeBucket can be updated with the latest details of
|
|
// the edge as they are announced on the network.
|
|
//
|
|
// maps: chanID -> pubKey1 || pubKey2 || restofEdgeInfo
|
|
edgeIndexBucket = []byte("edge-index")
|
|
|
|
// edgeUpdateIndexBucket is a sub-bucket of the main edgeBucket. This
|
|
// bucket contains an index which allows us to gauge the "freshness" of
|
|
// a channel's last updates.
|
|
//
|
|
// maps: updateTime || chanID -> nil
|
|
edgeUpdateIndexBucket = []byte("edge-update-index")
|
|
|
|
// channelPointBucket maps a channel's full outpoint (txid:index) to
|
|
// its short 8-byte channel ID. This bucket resides within the
|
|
// edgeBucket above, and can be used to quickly remove an edge due to
|
|
// the outpoint being spent, or to query for existence of a channel.
|
|
//
|
|
// maps: outPoint -> chanID
|
|
channelPointBucket = []byte("chan-index")
|
|
|
|
// graphMetaBucket is a top-level bucket which stores various meta-deta
|
|
// related to the on-disk channel graph. Data stored in this bucket
|
|
// includes the block to which the graph has been synced to, the total
|
|
// number of channels, etc.
|
|
graphMetaBucket = []byte("graph-meta")
|
|
|
|
// pruneLogBucket is a bucket within the graphMetaBucket that stores
|
|
// a mapping from the block height to the hash for the blocks used to
|
|
// prune the graph.
|
|
// Once a new block is discovered, any channels that have been closed
|
|
// (by spending the outpoint) can safely be removed from the graph, and
|
|
// the block is added to the prune log. We need to keep such a log for
|
|
// the case where a reorg happens, and we must "rewind" the state of the
|
|
// graph by removing channels that were previously confirmed. In such a
|
|
// case we'll remove all entries from the prune log with a block height
|
|
// that no longer exists.
|
|
pruneLogBucket = []byte("prune-log")
|
|
|
|
edgeBloomKey = []byte("edge-bloom")
|
|
nodeBloomKey = []byte("node-bloom")
|
|
)
|
|
|
|
// ChannelGraph is a persistent, on-disk graph representation of the Lightning
|
|
// Network. This struct can be used to implement path finding algorithms on top
|
|
// of, and also to update a node's view based on information received from the
|
|
// p2p network. Internally, the graph is stored using a modified adjacency list
|
|
// representation with some added object interaction possible with each
|
|
// serialized edge/node. The graph is stored is directed, meaning that are two
|
|
// edges stored for each channel: an inbound/outbound edge for each node pair.
|
|
// Nodes, edges, and edge information can all be added to the graph
|
|
// independently. Edge removal results in the deletion of all edge information
|
|
// for that edge.
|
|
type ChannelGraph struct {
|
|
db *DB
|
|
|
|
// TODO(roasbeef): store and update bloom filter to reduce disk access
|
|
// due to current gossip model
|
|
// * LRU cache for edges?
|
|
}
|
|
|
|
// Database returns a pointer to the underlying database.
|
|
func (c *ChannelGraph) Database() *DB {
|
|
return c.db
|
|
}
|
|
|
|
// ForEachChannel iterates through all the channel edges stored within the
|
|
// graph and invokes the passed callback for each edge. The callback takes two
|
|
// edges as since this is a directed graph, both the in/out edges are visited.
|
|
// If the callback returns an error, then the transaction is aborted and the
|
|
// iteration stops early.
|
|
//
|
|
// NOTE: If an edge can't be found, or wasn't advertised, then a nil pointer
|
|
// for that particular channel edge routing policy will be passed into the
|
|
// callback.
|
|
func (c *ChannelGraph) ForEachChannel(cb func(*ChannelEdgeInfo, *ChannelEdgePolicy, *ChannelEdgePolicy) error) error {
|
|
// TODO(roasbeef): ptr map to reduce # of allocs? no duplicates
|
|
|
|
return c.db.View(func(tx *bolt.Tx) error {
|
|
// First, grab the node bucket. This will be used to populate
|
|
// the Node pointers in each edge read from disk.
|
|
nodes := tx.Bucket(nodeBucket)
|
|
if nodes == nil {
|
|
return ErrGraphNotFound
|
|
}
|
|
|
|
// Next, grab the edge bucket which stores the edges, and also
|
|
// the index itself so we can group the directed edges together
|
|
// logically.
|
|
edges := tx.Bucket(edgeBucket)
|
|
if edges == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
edgeIndex := edges.Bucket(edgeIndexBucket)
|
|
if edgeIndex == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
|
|
// For each edge pair within the edge index, we fetch each edge
|
|
// itself and also the node information in order to fully
|
|
// populated the object.
|
|
return edgeIndex.ForEach(func(chanID, edgeInfoBytes []byte) error {
|
|
infoReader := bytes.NewReader(edgeInfoBytes)
|
|
edgeInfo, err := deserializeChanEdgeInfo(infoReader)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
edge1, edge2, err := fetchChanEdgePolicies(
|
|
edgeIndex, edges, nodes, chanID, c.db,
|
|
)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
// With both edges read, execute the call back. IF this
|
|
// function returns an error then the transaction will
|
|
// be aborted.
|
|
return cb(&edgeInfo, edge1, edge2)
|
|
})
|
|
})
|
|
}
|
|
|
|
// ForEachNode iterates through all the stored vertices/nodes in the graph,
|
|
// executing the passed callback with each node encountered. If the callback
|
|
// returns an error, then the transaction is aborted and the iteration stops
|
|
// early.
|
|
//
|
|
// If the caller wishes to re-use an existing boltdb transaction, then it
|
|
// should be passed as the first argument. Otherwise the first argument should
|
|
// be nil and a fresh transaction will be created to execute the graph
|
|
// traversal
|
|
//
|
|
// TODO(roasbeef): add iterator interface to allow for memory efficient graph
|
|
// traversal when graph gets mega
|
|
func (c *ChannelGraph) ForEachNode(tx *bolt.Tx, cb func(*bolt.Tx, *LightningNode) error) error {
|
|
traversal := func(tx *bolt.Tx) error {
|
|
// First grab the nodes bucket which stores the mapping from
|
|
// pubKey to node information.
|
|
nodes := tx.Bucket(nodeBucket)
|
|
if nodes == nil {
|
|
return ErrGraphNotFound
|
|
}
|
|
|
|
return nodes.ForEach(func(pubKey, nodeBytes []byte) error {
|
|
// If this is the source key, then we skip this
|
|
// iteration as the value for this key is a pubKey
|
|
// rather than raw node information.
|
|
if bytes.Equal(pubKey, sourceKey) || len(pubKey) != 33 {
|
|
return nil
|
|
}
|
|
|
|
nodeReader := bytes.NewReader(nodeBytes)
|
|
node, err := deserializeLightningNode(nodeReader)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
node.db = c.db
|
|
|
|
// Execute the callback, the transaction will abort if
|
|
// this returns an error.
|
|
return cb(tx, &node)
|
|
})
|
|
}
|
|
|
|
// If no transaction was provided, then we'll create a new transaction
|
|
// to execute the transaction within.
|
|
if tx == nil {
|
|
return c.db.View(traversal)
|
|
}
|
|
|
|
// Otherwise, we re-use the existing transaction to execute the graph
|
|
// traversal.
|
|
return traversal(tx)
|
|
}
|
|
|
|
// SourceNode returns the source node of the graph. The source node is treated
|
|
// as the center node within a star-graph. This method may be used to kick off
|
|
// a path finding algorithm in order to explore the reachability of another
|
|
// node based off the source node.
|
|
func (c *ChannelGraph) SourceNode() (*LightningNode, error) {
|
|
var source *LightningNode
|
|
err := c.db.View(func(tx *bolt.Tx) error {
|
|
node, err := c.sourceNode(tx)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
source = node
|
|
|
|
return nil
|
|
})
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
return source, nil
|
|
}
|
|
|
|
// sourceNode uses an existing database transaction and returns the source node
|
|
// of the graph. The source node is treated as the center node within a
|
|
// star-graph. This method may be used to kick off a path finding algorithm in
|
|
// order to explore the reachability of another node based off the source node.
|
|
func (c *ChannelGraph) sourceNode(tx *bolt.Tx) (*LightningNode, error) {
|
|
// First grab the nodes bucket which stores the mapping from
|
|
// pubKey to node information.
|
|
nodes := tx.Bucket(nodeBucket)
|
|
if nodes == nil {
|
|
return nil, ErrGraphNotFound
|
|
}
|
|
|
|
selfPub := nodes.Get(sourceKey)
|
|
if selfPub == nil {
|
|
return nil, ErrSourceNodeNotSet
|
|
}
|
|
|
|
// With the pubKey of the source node retrieved, we're able to
|
|
// fetch the full node information.
|
|
node, err := fetchLightningNode(nodes, selfPub)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
node.db = c.db
|
|
|
|
return &node, nil
|
|
}
|
|
|
|
// SetSourceNode sets the source node within the graph database. The source
|
|
// node is to be used as the center of a star-graph within path finding
|
|
// algorithms.
|
|
func (c *ChannelGraph) SetSourceNode(node *LightningNode) error {
|
|
nodePubBytes := node.PubKeyBytes[:]
|
|
|
|
return c.db.Update(func(tx *bolt.Tx) error {
|
|
// First grab the nodes bucket which stores the mapping from
|
|
// pubKey to node information.
|
|
nodes, err := tx.CreateBucketIfNotExists(nodeBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
// Next we create the mapping from source to the targeted
|
|
// public key.
|
|
if err := nodes.Put(sourceKey, nodePubBytes); err != nil {
|
|
return err
|
|
}
|
|
|
|
// Finally, we commit the information of the lightning node
|
|
// itself.
|
|
return addLightningNode(tx, node)
|
|
})
|
|
}
|
|
|
|
// AddLightningNode adds a vertex/node to the graph database. If the node is not
|
|
// in the database from before, this will add a new, unconnected one to the
|
|
// graph. If it is present from before, this will update that node's
|
|
// information. Note that this method is expected to only be called to update
|
|
// an already present node from a node announcement, or to insert a node found
|
|
// in a channel update.
|
|
//
|
|
// TODO(roasbeef): also need sig of announcement
|
|
func (c *ChannelGraph) AddLightningNode(node *LightningNode) error {
|
|
return c.db.Update(func(tx *bolt.Tx) error {
|
|
return addLightningNode(tx, node)
|
|
})
|
|
}
|
|
|
|
func addLightningNode(tx *bolt.Tx, node *LightningNode) error {
|
|
nodes, err := tx.CreateBucketIfNotExists(nodeBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
aliases, err := nodes.CreateBucketIfNotExists(aliasIndexBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
updateIndex, err := nodes.CreateBucketIfNotExists(
|
|
nodeUpdateIndexBucket,
|
|
)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
return putLightningNode(nodes, aliases, updateIndex, node)
|
|
}
|
|
|
|
// LookupAlias attempts to return the alias as advertised by the target node.
|
|
// TODO(roasbeef): currently assumes that aliases are unique...
|
|
func (c *ChannelGraph) LookupAlias(pub *btcec.PublicKey) (string, error) {
|
|
var alias string
|
|
|
|
err := c.db.View(func(tx *bolt.Tx) error {
|
|
nodes := tx.Bucket(nodeBucket)
|
|
if nodes == nil {
|
|
return ErrGraphNodesNotFound
|
|
}
|
|
|
|
aliases := nodes.Bucket(aliasIndexBucket)
|
|
if aliases == nil {
|
|
return ErrGraphNodesNotFound
|
|
}
|
|
|
|
nodePub := pub.SerializeCompressed()
|
|
a := aliases.Get(nodePub)
|
|
if a == nil {
|
|
return ErrNodeAliasNotFound
|
|
}
|
|
|
|
// TODO(roasbeef): should actually be using the utf-8
|
|
// package...
|
|
alias = string(a)
|
|
return nil
|
|
})
|
|
if err != nil {
|
|
return "", err
|
|
}
|
|
|
|
return alias, nil
|
|
}
|
|
|
|
// DeleteLightningNode starts a new database transaction to remove a vertex/node
|
|
// from the database according to the node's public key.
|
|
func (c *ChannelGraph) DeleteLightningNode(nodePub *btcec.PublicKey) error {
|
|
// TODO(roasbeef): ensure dangling edges are removed...
|
|
return c.db.Update(func(tx *bolt.Tx) error {
|
|
return c.deleteLightningNode(tx, nodePub.SerializeCompressed())
|
|
})
|
|
}
|
|
|
|
// deleteLightningNode uses an existing database transaction to remove a
|
|
// vertex/node from the database according to the node's public key.
|
|
func (c *ChannelGraph) deleteLightningNode(tx *bolt.Tx,
|
|
compressedPubKey []byte) error {
|
|
|
|
nodes := tx.Bucket(nodeBucket)
|
|
if nodes == nil {
|
|
return ErrGraphNodesNotFound
|
|
}
|
|
|
|
aliases := nodes.Bucket(aliasIndexBucket)
|
|
if aliases == nil {
|
|
return ErrGraphNodesNotFound
|
|
}
|
|
|
|
if err := aliases.Delete(compressedPubKey); err != nil {
|
|
return err
|
|
}
|
|
return nodes.Delete(compressedPubKey)
|
|
}
|
|
|
|
// AddChannelEdge adds a new (undirected, blank) edge to the graph database. An
|
|
// undirected edge from the two target nodes are created. The information
|
|
// stored denotes the static attributes of the channel, such as the channelID,
|
|
// the keys involved in creation of the channel, and the set of features that
|
|
// the channel supports. The chanPoint and chanID are used to uniquely identify
|
|
// the edge globally within the database.
|
|
func (c *ChannelGraph) AddChannelEdge(edge *ChannelEdgeInfo) error {
|
|
// Construct the channel's primary key which is the 8-byte channel ID.
|
|
var chanKey [8]byte
|
|
binary.BigEndian.PutUint64(chanKey[:], edge.ChannelID)
|
|
|
|
return c.db.Update(func(tx *bolt.Tx) error {
|
|
nodes, err := tx.CreateBucketIfNotExists(nodeBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
edges, err := tx.CreateBucketIfNotExists(edgeBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
edgeIndex, err := edges.CreateBucketIfNotExists(edgeIndexBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
chanIndex, err := edges.CreateBucketIfNotExists(channelPointBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
// First, attempt to check if this edge has already been
|
|
// created. If so, then we can exit early as this method is
|
|
// meant to be idempotent.
|
|
if edgeInfo := edgeIndex.Get(chanKey[:]); edgeInfo != nil {
|
|
return ErrEdgeAlreadyExist
|
|
}
|
|
|
|
// Before we insert the channel into the database, we'll ensure
|
|
// that both nodes already exist in the channel graph. If
|
|
// either node doesn't, then we'll insert a "shell" node that
|
|
// just includes its public key, so subsequent validation and
|
|
// queries can work properly.
|
|
_, node1Err := fetchLightningNode(nodes, edge.NodeKey1Bytes[:])
|
|
switch {
|
|
case node1Err == ErrGraphNodeNotFound:
|
|
node1Shell := LightningNode{
|
|
PubKeyBytes: edge.NodeKey1Bytes,
|
|
HaveNodeAnnouncement: false,
|
|
}
|
|
err := addLightningNode(tx, &node1Shell)
|
|
if err != nil {
|
|
return fmt.Errorf("unable to create shell node "+
|
|
"for: %x", edge.NodeKey1Bytes)
|
|
|
|
}
|
|
case node1Err != nil:
|
|
return err
|
|
}
|
|
|
|
_, node2Err := fetchLightningNode(nodes, edge.NodeKey2Bytes[:])
|
|
switch {
|
|
case node2Err == ErrGraphNodeNotFound:
|
|
node2Shell := LightningNode{
|
|
PubKeyBytes: edge.NodeKey2Bytes,
|
|
HaveNodeAnnouncement: false,
|
|
}
|
|
err := addLightningNode(tx, &node2Shell)
|
|
if err != nil {
|
|
return fmt.Errorf("unable to create shell node "+
|
|
"for: %x", edge.NodeKey2Bytes)
|
|
|
|
}
|
|
case node2Err != nil:
|
|
return err
|
|
}
|
|
|
|
// If the edge hasn't been created yet, then we'll first add it
|
|
// to the edge index in order to associate the edge between two
|
|
// nodes and also store the static components of the channel.
|
|
if err := putChanEdgeInfo(edgeIndex, edge, chanKey); err != nil {
|
|
return err
|
|
}
|
|
|
|
// Mark edge policies for both sides as unknown. This is to
|
|
// enable efficient incoming channel lookup for a node.
|
|
for _, key := range []*[33]byte{&edge.NodeKey1Bytes,
|
|
&edge.NodeKey2Bytes} {
|
|
|
|
err := putChanEdgePolicyUnknown(edges, edge.ChannelID,
|
|
key[:])
|
|
if err != nil {
|
|
return err
|
|
}
|
|
}
|
|
|
|
// Finally we add it to the channel index which maps channel
|
|
// points (outpoints) to the shorter channel ID's.
|
|
var b bytes.Buffer
|
|
if err := writeOutpoint(&b, &edge.ChannelPoint); err != nil {
|
|
return err
|
|
}
|
|
return chanIndex.Put(b.Bytes(), chanKey[:])
|
|
})
|
|
}
|
|
|
|
// HasChannelEdge returns true if the database knows of a channel edge with the
|
|
// passed channel ID, and false otherwise. If an edge with that ID is found
|
|
// within the graph, then two time stamps representing the last time the edge
|
|
// was updated for both directed edges are returned along with the boolean.
|
|
func (c *ChannelGraph) HasChannelEdge(chanID uint64) (time.Time, time.Time, bool, error) {
|
|
// TODO(roasbeef): check internal bloom filter first
|
|
|
|
var (
|
|
node1UpdateTime time.Time
|
|
node2UpdateTime time.Time
|
|
exists bool
|
|
)
|
|
|
|
if err := c.db.View(func(tx *bolt.Tx) error {
|
|
edges := tx.Bucket(edgeBucket)
|
|
if edges == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
edgeIndex := edges.Bucket(edgeIndexBucket)
|
|
if edgeIndex == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
|
|
var channelID [8]byte
|
|
byteOrder.PutUint64(channelID[:], chanID)
|
|
if edgeIndex.Get(channelID[:]) == nil {
|
|
exists = false
|
|
return nil
|
|
}
|
|
|
|
exists = true
|
|
|
|
// If the channel has been found in the graph, then retrieve
|
|
// the edges itself so we can return the last updated
|
|
// timestamps.
|
|
nodes := tx.Bucket(nodeBucket)
|
|
if nodes == nil {
|
|
return ErrGraphNodeNotFound
|
|
}
|
|
|
|
e1, e2, err := fetchChanEdgePolicies(edgeIndex, edges, nodes,
|
|
channelID[:], c.db)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
// As we may have only one of the edges populated, only set the
|
|
// update time if the edge was found in the database.
|
|
if e1 != nil {
|
|
node1UpdateTime = e1.LastUpdate
|
|
}
|
|
if e2 != nil {
|
|
node2UpdateTime = e2.LastUpdate
|
|
}
|
|
|
|
return nil
|
|
}); err != nil {
|
|
return time.Time{}, time.Time{}, exists, err
|
|
}
|
|
|
|
return node1UpdateTime, node2UpdateTime, exists, nil
|
|
}
|
|
|
|
// UpdateChannelEdge retrieves and update edge of the graph database. Method
|
|
// only reserved for updating an edge info after its already been created.
|
|
// In order to maintain this constraints, we return an error in the scenario
|
|
// that an edge info hasn't yet been created yet, but someone attempts to update
|
|
// it.
|
|
func (c *ChannelGraph) UpdateChannelEdge(edge *ChannelEdgeInfo) error {
|
|
// Construct the channel's primary key which is the 8-byte channel ID.
|
|
var chanKey [8]byte
|
|
binary.BigEndian.PutUint64(chanKey[:], edge.ChannelID)
|
|
|
|
return c.db.Update(func(tx *bolt.Tx) error {
|
|
edges, err := tx.CreateBucketIfNotExists(edgeBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
edgeIndex, err := edges.CreateBucketIfNotExists(edgeIndexBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
if edgeInfo := edgeIndex.Get(chanKey[:]); edgeInfo == nil {
|
|
return ErrEdgeNotFound
|
|
}
|
|
|
|
return putChanEdgeInfo(edgeIndex, edge, chanKey)
|
|
})
|
|
}
|
|
|
|
const (
|
|
// pruneTipBytes is the total size of the value which stores a prune
|
|
// entry of the graph in the prune log. The "prune tip" is the last
|
|
// entry in the prune log, and indicates if the channel graph is in
|
|
// sync with the current UTXO state. The structure of the value
|
|
// is: blockHash, taking 32 bytes total.
|
|
pruneTipBytes = 32
|
|
)
|
|
|
|
// PruneGraph prunes newly closed channels from the channel graph in response
|
|
// to a new block being solved on the network. Any transactions which spend the
|
|
// funding output of any known channels within he graph will be deleted.
|
|
// Additionally, the "prune tip", or the last block which has been used to
|
|
// prune the graph is stored so callers can ensure the graph is fully in sync
|
|
// with the current UTXO state. A slice of channels that have been closed by
|
|
// the target block are returned if the function succeeds without error.
|
|
func (c *ChannelGraph) PruneGraph(spentOutputs []*wire.OutPoint,
|
|
blockHash *chainhash.Hash, blockHeight uint32) ([]*ChannelEdgeInfo, error) {
|
|
|
|
var chansClosed []*ChannelEdgeInfo
|
|
|
|
err := c.db.Update(func(tx *bolt.Tx) error {
|
|
// First grab the edges bucket which houses the information
|
|
// we'd like to delete
|
|
edges, err := tx.CreateBucketIfNotExists(edgeBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
// Next grab the two edge indexes which will also need to be updated.
|
|
edgeIndex, err := edges.CreateBucketIfNotExists(edgeIndexBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
chanIndex, err := edges.CreateBucketIfNotExists(channelPointBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
nodes, err := tx.CreateBucketIfNotExists(nodeBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
// For each of the outpoints that have been spent within the
|
|
// block, we attempt to delete them from the graph as if that
|
|
// outpoint was a channel, then it has now been closed.
|
|
for _, chanPoint := range spentOutputs {
|
|
// TODO(roasbeef): load channel bloom filter, continue
|
|
// if NOT if filter
|
|
|
|
var opBytes bytes.Buffer
|
|
if err := writeOutpoint(&opBytes, chanPoint); err != nil {
|
|
return nil
|
|
}
|
|
|
|
// First attempt to see if the channel exists within
|
|
// the database, if not, then we can exit early.
|
|
chanID := chanIndex.Get(opBytes.Bytes())
|
|
if chanID == nil {
|
|
continue
|
|
}
|
|
|
|
// However, if it does, then we'll read out the full
|
|
// version so we can add it to the set of deleted
|
|
// channels.
|
|
edgeInfo, err := fetchChanEdgeInfo(edgeIndex, chanID)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
// Attempt to delete the channel, an ErrEdgeNotFound
|
|
// will be returned if that outpoint isn't known to be
|
|
// a channel. If no error is returned, then a channel
|
|
// was successfully pruned.
|
|
err = delChannelByEdge(
|
|
edges, edgeIndex, chanIndex, nodes, chanPoint,
|
|
)
|
|
if err != nil && err != ErrEdgeNotFound {
|
|
return err
|
|
}
|
|
|
|
chansClosed = append(chansClosed, &edgeInfo)
|
|
}
|
|
|
|
metaBucket, err := tx.CreateBucketIfNotExists(graphMetaBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
pruneBucket, err := metaBucket.CreateBucketIfNotExists(pruneLogBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
// With the graph pruned, add a new entry to the prune log,
|
|
// which can be used to check if the graph is fully synced with
|
|
// the current UTXO state.
|
|
var blockHeightBytes [4]byte
|
|
byteOrder.PutUint32(blockHeightBytes[:], blockHeight)
|
|
|
|
var newTip [pruneTipBytes]byte
|
|
copy(newTip[:], blockHash[:])
|
|
|
|
err = pruneBucket.Put(blockHeightBytes[:], newTip[:])
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
// Now that the graph has been pruned, we'll also attempt to
|
|
// prune any nodes that have had a channel closed within the
|
|
// latest block.
|
|
return c.pruneGraphNodes(tx, nodes, edgeIndex)
|
|
})
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
return chansClosed, nil
|
|
}
|
|
|
|
// PruneGraphNodes is a garbage collection method which attempts to prune out
|
|
// any nodes from the channel graph that are currently unconnected. This ensure
|
|
// that we only maintain a graph of reachable nodes. In the event that a pruned
|
|
// node gains more channels, it will be re-added back to the graph.
|
|
func (c *ChannelGraph) PruneGraphNodes() error {
|
|
return c.db.Update(func(tx *bolt.Tx) error {
|
|
nodes, err := tx.CreateBucketIfNotExists(nodeBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
edges := tx.Bucket(edgeBucket)
|
|
if edges == nil {
|
|
return ErrGraphNotFound
|
|
}
|
|
edgeIndex := edges.Bucket(edgeIndexBucket)
|
|
if edgeIndex == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
|
|
return c.pruneGraphNodes(tx, nodes, edgeIndex)
|
|
})
|
|
}
|
|
|
|
// pruneGraphNodes attempts to remove any nodes from the graph who have had a
|
|
// channel closed within the current block. If the node still has existing
|
|
// channels in the graph, this will act as a no-op.
|
|
func (c *ChannelGraph) pruneGraphNodes(tx *bolt.Tx, nodes *bolt.Bucket,
|
|
edgeIndex *bolt.Bucket) error {
|
|
|
|
log.Trace("Pruning nodes from graph with no open channels")
|
|
|
|
// We'll retrieve the graph's source node to ensure we don't remove it
|
|
// even if it no longer has any open channels.
|
|
sourceNode, err := c.sourceNode(tx)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
// We'll use this map to keep count the number of references to a node
|
|
// in the graph. A node should only be removed once it has no more
|
|
// references in the graph.
|
|
nodeRefCounts := make(map[[33]byte]int)
|
|
err = nodes.ForEach(func(pubKey, nodeBytes []byte) error {
|
|
// If this is the source key, then we skip this
|
|
// iteration as the value for this key is a pubKey
|
|
// rather than raw node information.
|
|
if bytes.Equal(pubKey, sourceKey) || len(pubKey) != 33 {
|
|
return nil
|
|
}
|
|
|
|
var nodePub [33]byte
|
|
copy(nodePub[:], pubKey)
|
|
nodeRefCounts[nodePub] = 0
|
|
|
|
return nil
|
|
})
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
// To ensure we never delete the source node, we'll start off by
|
|
// bumping its ref count to 1.
|
|
nodeRefCounts[sourceNode.PubKeyBytes] = 1
|
|
|
|
// Next, we'll run through the edgeIndex which maps a channel ID to the
|
|
// edge info. We'll use this scan to populate our reference count map
|
|
// above.
|
|
err = edgeIndex.ForEach(func(chanID, edgeInfoBytes []byte) error {
|
|
// The first 66 bytes of the edge info contain the pubkeys of
|
|
// the nodes that this edge attaches. We'll extract them, and
|
|
// add them to the ref count map.
|
|
var node1, node2 [33]byte
|
|
copy(node1[:], edgeInfoBytes[:33])
|
|
copy(node2[:], edgeInfoBytes[33:])
|
|
|
|
// With the nodes extracted, we'll increase the ref count of
|
|
// each of the nodes.
|
|
nodeRefCounts[node1]++
|
|
nodeRefCounts[node2]++
|
|
|
|
return nil
|
|
})
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
// Finally, we'll make a second pass over the set of nodes, and delete
|
|
// any nodes that have a ref count of zero.
|
|
var numNodesPruned int
|
|
for nodePubKey, refCount := range nodeRefCounts {
|
|
// If the ref count of the node isn't zero, then we can safely
|
|
// skip it as it still has edges to or from it within the
|
|
// graph.
|
|
if refCount != 0 {
|
|
continue
|
|
}
|
|
|
|
// If we reach this point, then there are no longer any edges
|
|
// that connect this node, so we can delete it.
|
|
if err := c.deleteLightningNode(tx, nodePubKey[:]); err != nil {
|
|
log.Warnf("Unable to prune node %x from the "+
|
|
"graph: %v", nodePubKey, err)
|
|
continue
|
|
}
|
|
|
|
log.Infof("Pruned unconnected node %x from channel graph",
|
|
nodePubKey[:])
|
|
|
|
numNodesPruned++
|
|
}
|
|
|
|
if numNodesPruned > 0 {
|
|
log.Infof("Pruned %v unconnected nodes from the channel graph",
|
|
numNodesPruned)
|
|
}
|
|
|
|
return nil
|
|
}
|
|
|
|
// DisconnectBlockAtHeight is used to indicate that the block specified
|
|
// by the passed height has been disconnected from the main chain. This
|
|
// will "rewind" the graph back to the height below, deleting channels
|
|
// that are no longer confirmed from the graph. The prune log will be
|
|
// set to the last prune height valid for the remaining chain.
|
|
// Channels that were removed from the graph resulting from the
|
|
// disconnected block are returned.
|
|
func (c *ChannelGraph) DisconnectBlockAtHeight(height uint32) ([]*ChannelEdgeInfo,
|
|
error) {
|
|
|
|
// Every channel having a ShortChannelID starting at 'height'
|
|
// will no longer be confirmed.
|
|
startShortChanID := lnwire.ShortChannelID{
|
|
BlockHeight: height,
|
|
}
|
|
|
|
// Delete everything after this height from the db.
|
|
endShortChanID := lnwire.ShortChannelID{
|
|
BlockHeight: math.MaxUint32 & 0x00ffffff,
|
|
TxIndex: math.MaxUint32 & 0x00ffffff,
|
|
TxPosition: math.MaxUint16,
|
|
}
|
|
// The block height will be the 3 first bytes of the channel IDs.
|
|
var chanIDStart [8]byte
|
|
byteOrder.PutUint64(chanIDStart[:], startShortChanID.ToUint64())
|
|
var chanIDEnd [8]byte
|
|
byteOrder.PutUint64(chanIDEnd[:], endShortChanID.ToUint64())
|
|
|
|
// Keep track of the channels that are removed from the graph.
|
|
var removedChans []*ChannelEdgeInfo
|
|
|
|
if err := c.db.Update(func(tx *bolt.Tx) error {
|
|
edges, err := tx.CreateBucketIfNotExists(edgeBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
edgeIndex, err := edges.CreateBucketIfNotExists(edgeIndexBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
chanIndex, err := edges.CreateBucketIfNotExists(channelPointBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
nodes, err := tx.CreateBucketIfNotExists(nodeBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
// Scan from chanIDStart to chanIDEnd, deleting every
|
|
// found edge.
|
|
cursor := edgeIndex.Cursor()
|
|
for k, v := cursor.Seek(chanIDStart[:]); k != nil &&
|
|
bytes.Compare(k, chanIDEnd[:]) <= 0; k, v = cursor.Next() {
|
|
|
|
edgeInfoReader := bytes.NewReader(v)
|
|
edgeInfo, err := deserializeChanEdgeInfo(edgeInfoReader)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
err = delChannelByEdge(
|
|
edges, edgeIndex, chanIndex, nodes, &edgeInfo.ChannelPoint,
|
|
)
|
|
if err != nil && err != ErrEdgeNotFound {
|
|
return err
|
|
}
|
|
|
|
removedChans = append(removedChans, &edgeInfo)
|
|
}
|
|
|
|
// Delete all the entries in the prune log having a height
|
|
// greater or equal to the block disconnected.
|
|
metaBucket, err := tx.CreateBucketIfNotExists(graphMetaBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
pruneBucket, err := metaBucket.CreateBucketIfNotExists(pruneLogBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
var pruneKeyStart [4]byte
|
|
byteOrder.PutUint32(pruneKeyStart[:], height)
|
|
|
|
var pruneKeyEnd [4]byte
|
|
byteOrder.PutUint32(pruneKeyEnd[:], math.MaxUint32)
|
|
|
|
pruneCursor := pruneBucket.Cursor()
|
|
for k, _ := pruneCursor.Seek(pruneKeyStart[:]); k != nil &&
|
|
bytes.Compare(k, pruneKeyEnd[:]) <= 0; k, _ = pruneCursor.Next() {
|
|
if err := pruneCursor.Delete(); err != nil {
|
|
return err
|
|
}
|
|
}
|
|
|
|
return nil
|
|
}); err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
return removedChans, nil
|
|
}
|
|
|
|
// PruneTip returns the block height and hash of the latest block that has been
|
|
// used to prune channels in the graph. Knowing the "prune tip" allows callers
|
|
// to tell if the graph is currently in sync with the current best known UTXO
|
|
// state.
|
|
func (c *ChannelGraph) PruneTip() (*chainhash.Hash, uint32, error) {
|
|
var (
|
|
tipHash chainhash.Hash
|
|
tipHeight uint32
|
|
)
|
|
|
|
err := c.db.View(func(tx *bolt.Tx) error {
|
|
graphMeta := tx.Bucket(graphMetaBucket)
|
|
if graphMeta == nil {
|
|
return ErrGraphNotFound
|
|
}
|
|
pruneBucket := graphMeta.Bucket(pruneLogBucket)
|
|
if pruneBucket == nil {
|
|
return ErrGraphNeverPruned
|
|
}
|
|
|
|
pruneCursor := pruneBucket.Cursor()
|
|
|
|
// The prune key with the largest block height will be our
|
|
// prune tip.
|
|
k, v := pruneCursor.Last()
|
|
if k == nil {
|
|
return ErrGraphNeverPruned
|
|
}
|
|
|
|
// Once we have the prune tip, the value will be the block hash,
|
|
// and the key the block height.
|
|
copy(tipHash[:], v[:])
|
|
tipHeight = byteOrder.Uint32(k[:])
|
|
|
|
return nil
|
|
})
|
|
if err != nil {
|
|
return nil, 0, err
|
|
}
|
|
|
|
return &tipHash, tipHeight, nil
|
|
}
|
|
|
|
// DeleteChannelEdge removes an edge from the database as identified by its
|
|
// funding outpoint. If the edge does not exist within the database, then
|
|
// ErrEdgeNotFound will be returned.
|
|
func (c *ChannelGraph) DeleteChannelEdge(chanPoint *wire.OutPoint) error {
|
|
// TODO(roasbeef): possibly delete from node bucket if node has no more
|
|
// channels
|
|
// TODO(roasbeef): don't delete both edges?
|
|
|
|
return c.db.Update(func(tx *bolt.Tx) error {
|
|
// First grab the edges bucket which houses the information
|
|
// we'd like to delete
|
|
edges, err := tx.CreateBucketIfNotExists(edgeBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
// Next grab the two edge indexes which will also need to be updated.
|
|
edgeIndex, err := edges.CreateBucketIfNotExists(edgeIndexBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
chanIndex, err := edges.CreateBucketIfNotExists(channelPointBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
nodes, err := tx.CreateBucketIfNotExists(nodeBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
return delChannelByEdge(edges, edgeIndex, chanIndex, nodes, chanPoint)
|
|
})
|
|
}
|
|
|
|
// ChannelID attempt to lookup the 8-byte compact channel ID which maps to the
|
|
// passed channel point (outpoint). If the passed channel doesn't exist within
|
|
// the database, then ErrEdgeNotFound is returned.
|
|
func (c *ChannelGraph) ChannelID(chanPoint *wire.OutPoint) (uint64, error) {
|
|
var chanID uint64
|
|
|
|
var b bytes.Buffer
|
|
if err := writeOutpoint(&b, chanPoint); err != nil {
|
|
return 0, nil
|
|
}
|
|
|
|
if err := c.db.View(func(tx *bolt.Tx) error {
|
|
edges := tx.Bucket(edgeBucket)
|
|
if edges == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
chanIndex := edges.Bucket(channelPointBucket)
|
|
if chanIndex == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
|
|
chanIDBytes := chanIndex.Get(b.Bytes())
|
|
if chanIDBytes == nil {
|
|
return ErrEdgeNotFound
|
|
}
|
|
|
|
chanID = byteOrder.Uint64(chanIDBytes)
|
|
|
|
return nil
|
|
}); err != nil {
|
|
return 0, err
|
|
}
|
|
|
|
return chanID, nil
|
|
}
|
|
|
|
// TODO(roasbeef): allow updates to use Batch?
|
|
|
|
// HighestChanID returns the "highest" known channel ID in the channel graph.
|
|
// This represents the "newest" channel from the PoV of the chain. This method
|
|
// can be used by peers to quickly determine if they're graphs are in sync.
|
|
func (c *ChannelGraph) HighestChanID() (uint64, error) {
|
|
var cid uint64
|
|
|
|
err := c.db.View(func(tx *bolt.Tx) error {
|
|
edges := tx.Bucket(edgeBucket)
|
|
if edges == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
edgeIndex := edges.Bucket(edgeIndexBucket)
|
|
if edgeIndex == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
|
|
// In order to find the highest chan ID, we'll fetch a cursor
|
|
// and use that to seek to the "end" of our known rage.
|
|
cidCursor := edgeIndex.Cursor()
|
|
|
|
lastChanID, _ := cidCursor.Last()
|
|
|
|
// If there's no key, then this means that we don't actually
|
|
// know of any channels, so we'll return a predicable error.
|
|
if lastChanID == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
|
|
// Otherwise, we'll de serialize the channel ID and return it
|
|
// to the caller.
|
|
cid = byteOrder.Uint64(lastChanID)
|
|
return nil
|
|
})
|
|
if err != nil && err != ErrGraphNoEdgesFound {
|
|
return 0, err
|
|
}
|
|
|
|
return cid, nil
|
|
}
|
|
|
|
// ChannelEdge represents the complete set of information for a channel edge in
|
|
// the known channel graph. This struct couples the core information of the
|
|
// edge as well as each of the known advertised edge policies.
|
|
type ChannelEdge struct {
|
|
// Info contains all the static information describing the channel.
|
|
Info *ChannelEdgeInfo
|
|
|
|
// Policy1 points to the "first" edge policy of the channel containing
|
|
// the dynamic information required to properly route through the edge.
|
|
Policy1 *ChannelEdgePolicy
|
|
|
|
// Policy2 points to the "second" edge policy of the channel containing
|
|
// the dynamic information required to properly route through the edge.
|
|
Policy2 *ChannelEdgePolicy
|
|
}
|
|
|
|
// ChanUpdatesInHorizon returns all the known channel edges which have at least
|
|
// one edge that has an update timestamp within the specified horizon.
|
|
func (c *ChannelGraph) ChanUpdatesInHorizon(startTime, endTime time.Time) ([]ChannelEdge, error) {
|
|
var edgesInHorizon []ChannelEdge
|
|
|
|
err := c.db.View(func(tx *bolt.Tx) error {
|
|
edges := tx.Bucket(edgeBucket)
|
|
if edges == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
edgeIndex := edges.Bucket(edgeIndexBucket)
|
|
if edgeIndex == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
edgeUpdateIndex := edges.Bucket(edgeUpdateIndexBucket)
|
|
if edgeUpdateIndex == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
|
|
nodes := tx.Bucket(nodeBucket)
|
|
if nodes == nil {
|
|
return ErrGraphNodesNotFound
|
|
}
|
|
|
|
// We'll now obtain a cursor to perform a range query within
|
|
// the index to find all channels within the horizon.
|
|
updateCursor := edgeUpdateIndex.Cursor()
|
|
|
|
var startTimeBytes, endTimeBytes [8 + 8]byte
|
|
byteOrder.PutUint64(
|
|
startTimeBytes[:8], uint64(startTime.Unix()),
|
|
)
|
|
byteOrder.PutUint64(
|
|
endTimeBytes[:8], uint64(endTime.Unix()),
|
|
)
|
|
|
|
// With our start and end times constructed, we'll step through
|
|
// the index collecting the info and policy of each update of
|
|
// each channel that has a last update within the time range.
|
|
for indexKey, _ := updateCursor.Seek(startTimeBytes[:]); indexKey != nil &&
|
|
bytes.Compare(indexKey, endTimeBytes[:]) <= 0; indexKey, _ = updateCursor.Next() {
|
|
|
|
// We have a new eligible entry, so we'll slice of the
|
|
// chan ID so we can query it in the DB.
|
|
chanID := indexKey[8:]
|
|
|
|
// First, we'll fetch the static edge information.
|
|
edgeInfo, err := fetchChanEdgeInfo(edgeIndex, chanID)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
// With the static information obtained, we'll now
|
|
// fetch the dynamic policy info.
|
|
edge1, edge2, err := fetchChanEdgePolicies(
|
|
edgeIndex, edges, nodes, chanID, c.db,
|
|
)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
// Finally, we'll collate this edge with the rest of
|
|
// edges to be returned.
|
|
edgesInHorizon = append(edgesInHorizon, ChannelEdge{
|
|
Info: &edgeInfo,
|
|
Policy1: edge1,
|
|
Policy2: edge2,
|
|
})
|
|
}
|
|
|
|
return nil
|
|
})
|
|
switch {
|
|
case err == ErrGraphNoEdgesFound:
|
|
fallthrough
|
|
case err == ErrGraphNodesNotFound:
|
|
break
|
|
|
|
case err != nil:
|
|
return nil, err
|
|
}
|
|
|
|
return edgesInHorizon, nil
|
|
}
|
|
|
|
// NodeUpdatesInHorizon returns all the known lightning node which have an
|
|
// update timestamp within the passed range. This method can be used by two
|
|
// nodes to quickly determine if they have the same set of up to date node
|
|
// announcements.
|
|
func (c *ChannelGraph) NodeUpdatesInHorizon(startTime, endTime time.Time) ([]LightningNode, error) {
|
|
var nodesInHorizon []LightningNode
|
|
|
|
err := c.db.View(func(tx *bolt.Tx) error {
|
|
nodes := tx.Bucket(nodeBucket)
|
|
if nodes == nil {
|
|
return ErrGraphNodesNotFound
|
|
}
|
|
|
|
nodeUpdateIndex := nodes.Bucket(nodeUpdateIndexBucket)
|
|
if nodeUpdateIndex == nil {
|
|
return ErrGraphNodesNotFound
|
|
}
|
|
|
|
// We'll now obtain a cursor to perform a range query within
|
|
// the index to find all node announcements within the horizon.
|
|
updateCursor := nodeUpdateIndex.Cursor()
|
|
|
|
var startTimeBytes, endTimeBytes [8 + 33]byte
|
|
byteOrder.PutUint64(
|
|
startTimeBytes[:8], uint64(startTime.Unix()),
|
|
)
|
|
byteOrder.PutUint64(
|
|
endTimeBytes[:8], uint64(endTime.Unix()),
|
|
)
|
|
|
|
// With our start and end times constructed, we'll step through
|
|
// the index collecting info for each node within the time
|
|
// range.
|
|
for indexKey, _ := updateCursor.Seek(startTimeBytes[:]); indexKey != nil &&
|
|
bytes.Compare(indexKey, endTimeBytes[:]) <= 0; indexKey, _ = updateCursor.Next() {
|
|
|
|
nodePub := indexKey[8:]
|
|
node, err := fetchLightningNode(nodes, nodePub)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
node.db = c.db
|
|
|
|
nodesInHorizon = append(nodesInHorizon, node)
|
|
}
|
|
|
|
return nil
|
|
})
|
|
switch {
|
|
case err == ErrGraphNoEdgesFound:
|
|
fallthrough
|
|
case err == ErrGraphNodesNotFound:
|
|
break
|
|
|
|
case err != nil:
|
|
return nil, err
|
|
}
|
|
|
|
return nodesInHorizon, nil
|
|
}
|
|
|
|
// FilterKnownChanIDs takes a set of channel IDs and return the subset of chan
|
|
// ID's that we don't know of in the passed set. In other words, we perform a
|
|
// set difference of our set of chan ID's and the ones passed in. This method
|
|
// can be used by callers to determine the set of channels ta peer knows of
|
|
// that we don't.
|
|
func (c *ChannelGraph) FilterKnownChanIDs(chanIDs []uint64) ([]uint64, error) {
|
|
var newChanIDs []uint64
|
|
|
|
err := c.db.View(func(tx *bolt.Tx) error {
|
|
edges := tx.Bucket(edgeBucket)
|
|
if edges == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
edgeIndex := edges.Bucket(edgeIndexBucket)
|
|
if edgeIndex == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
|
|
// We'll run through the set of chanIDs and collate only the
|
|
// set of channel that are unable to be found within our db.
|
|
var cidBytes [8]byte
|
|
for _, cid := range chanIDs {
|
|
byteOrder.PutUint64(cidBytes[:], cid)
|
|
|
|
if v := edgeIndex.Get(cidBytes[:]); v == nil {
|
|
newChanIDs = append(newChanIDs, cid)
|
|
}
|
|
}
|
|
|
|
return nil
|
|
})
|
|
switch {
|
|
// If we don't know of any edges yet, then we'll return the entire set
|
|
// of chan IDs specified.
|
|
case err == ErrGraphNoEdgesFound:
|
|
return chanIDs, nil
|
|
|
|
case err != nil:
|
|
return nil, err
|
|
}
|
|
|
|
return newChanIDs, nil
|
|
}
|
|
|
|
// FilterChannelRange returns the channel ID's of all known channels which were
|
|
// mined in a block height within the passed range. This method can be used to
|
|
// quickly share with a peer the set of channels we know of within a particular
|
|
// range to catch them up after a period of time offline.
|
|
func (c *ChannelGraph) FilterChannelRange(startHeight, endHeight uint32) ([]uint64, error) {
|
|
var chanIDs []uint64
|
|
|
|
startChanID := &lnwire.ShortChannelID{
|
|
BlockHeight: startHeight,
|
|
}
|
|
|
|
endChanID := lnwire.ShortChannelID{
|
|
BlockHeight: endHeight,
|
|
TxIndex: math.MaxUint32 & 0x00ffffff,
|
|
TxPosition: math.MaxUint16,
|
|
}
|
|
|
|
// As we need to perform a range scan, we'll convert the starting and
|
|
// ending height to their corresponding values when encoded using short
|
|
// channel ID's.
|
|
var chanIDStart, chanIDEnd [8]byte
|
|
byteOrder.PutUint64(chanIDStart[:], startChanID.ToUint64())
|
|
byteOrder.PutUint64(chanIDEnd[:], endChanID.ToUint64())
|
|
|
|
err := c.db.View(func(tx *bolt.Tx) error {
|
|
edges := tx.Bucket(edgeBucket)
|
|
if edges == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
edgeIndex := edges.Bucket(edgeIndexBucket)
|
|
if edgeIndex == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
|
|
cursor := edgeIndex.Cursor()
|
|
|
|
// We'll now iterate through the database, and find each
|
|
// channel ID that resides within the specified range.
|
|
var cid uint64
|
|
for k, _ := cursor.Seek(chanIDStart[:]); k != nil &&
|
|
bytes.Compare(k, chanIDEnd[:]) <= 0; k, _ = cursor.Next() {
|
|
|
|
// This channel ID rests within the target range, so
|
|
// we'll convert it into an integer and add it to our
|
|
// returned set.
|
|
cid = byteOrder.Uint64(k)
|
|
chanIDs = append(chanIDs, cid)
|
|
}
|
|
|
|
return nil
|
|
})
|
|
switch {
|
|
// If we don't know of any channels yet, then there's nothing to
|
|
// filter, so we'll return an empty slice.
|
|
case err == ErrGraphNoEdgesFound:
|
|
return chanIDs, nil
|
|
|
|
case err != nil:
|
|
return nil, err
|
|
}
|
|
|
|
return chanIDs, nil
|
|
}
|
|
|
|
// FetchChanInfos returns the set of channel edges that correspond to the
|
|
// passed channel ID's. This can be used to respond to peer queries that are
|
|
// seeking to fill in gaps in their view of the channel graph.
|
|
func (c *ChannelGraph) FetchChanInfos(chanIDs []uint64) ([]ChannelEdge, error) {
|
|
// TODO(roasbeef): sort cids?
|
|
|
|
var (
|
|
chanEdges []ChannelEdge
|
|
cidBytes [8]byte
|
|
)
|
|
|
|
err := c.db.View(func(tx *bolt.Tx) error {
|
|
edges := tx.Bucket(edgeBucket)
|
|
if edges == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
edgeIndex := edges.Bucket(edgeIndexBucket)
|
|
if edgeIndex == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
nodes := tx.Bucket(nodeBucket)
|
|
if nodes == nil {
|
|
return ErrGraphNotFound
|
|
}
|
|
|
|
for _, cid := range chanIDs {
|
|
byteOrder.PutUint64(cidBytes[:], cid)
|
|
|
|
// First, we'll fetch the static edge information.
|
|
edgeInfo, err := fetchChanEdgeInfo(
|
|
edgeIndex, cidBytes[:],
|
|
)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
// With the static information obtained, we'll now
|
|
// fetch the dynamic policy info.
|
|
edge1, edge2, err := fetchChanEdgePolicies(
|
|
edgeIndex, edges, nodes, cidBytes[:], c.db,
|
|
)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
chanEdges = append(chanEdges, ChannelEdge{
|
|
Info: &edgeInfo,
|
|
Policy1: edge1,
|
|
Policy2: edge2,
|
|
})
|
|
}
|
|
return nil
|
|
})
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
return chanEdges, nil
|
|
}
|
|
|
|
func delEdgeUpdateIndexEntry(edgesBucket *bolt.Bucket, chanID uint64,
|
|
edge1, edge2 *ChannelEdgePolicy) error {
|
|
|
|
// First, we'll fetch the edge update index bucket which currently
|
|
// stores an entry for the channel we're about to delete.
|
|
updateIndex, err := edgesBucket.CreateBucketIfNotExists(
|
|
edgeUpdateIndexBucket,
|
|
)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
// Now that we have the bucket, we'll attempt to construct a template
|
|
// for the index key: updateTime || chanid.
|
|
var indexKey [8 + 8]byte
|
|
byteOrder.PutUint64(indexKey[8:], chanID)
|
|
|
|
// With the template constructed, we'll attempt to delete an entry that
|
|
// would have been created by both edges: we'll alternate the update
|
|
// times, as one may had overridden the other.
|
|
if edge1 != nil {
|
|
byteOrder.PutUint64(indexKey[:8], uint64(edge1.LastUpdate.Unix()))
|
|
if err := updateIndex.Delete(indexKey[:]); err != nil {
|
|
return err
|
|
}
|
|
}
|
|
|
|
// We'll also attempt to delete the entry that may have been created by
|
|
// the second edge.
|
|
if edge2 != nil {
|
|
byteOrder.PutUint64(indexKey[:8], uint64(edge2.LastUpdate.Unix()))
|
|
if err := updateIndex.Delete(indexKey[:]); err != nil {
|
|
return err
|
|
}
|
|
}
|
|
|
|
return nil
|
|
}
|
|
|
|
func delChannelByEdge(edges *bolt.Bucket, edgeIndex *bolt.Bucket,
|
|
chanIndex *bolt.Bucket, nodes *bolt.Bucket, chanPoint *wire.OutPoint) error {
|
|
var b bytes.Buffer
|
|
if err := writeOutpoint(&b, chanPoint); err != nil {
|
|
return err
|
|
}
|
|
|
|
// If the channel's outpoint doesn't exist within the outpoint index,
|
|
// then the edge does not exist.
|
|
chanID := chanIndex.Get(b.Bytes())
|
|
if chanID == nil {
|
|
return ErrEdgeNotFound
|
|
}
|
|
|
|
// Otherwise we obtain the two public keys from the mapping: chanID ->
|
|
// pubKey1 || pubKey2. With this, we can construct the keys which house
|
|
// both of the directed edges for this channel.
|
|
nodeKeys := edgeIndex.Get(chanID)
|
|
if nodeKeys == nil {
|
|
return fmt.Errorf("could not find nodekeys for chanID %v",
|
|
chanID)
|
|
}
|
|
|
|
// We'll also remove the entry in the edge update index bucket before
|
|
// we delete the edges themselves so we can access their last update
|
|
// times.
|
|
cid := byteOrder.Uint64(chanID)
|
|
edge1, edge2, err := fetchChanEdgePolicies(
|
|
edgeIndex, edges, nodes, chanID, nil,
|
|
)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
err = delEdgeUpdateIndexEntry(edges, cid, edge1, edge2)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
// The edge key is of the format pubKey || chanID. First we construct
|
|
// the latter half, populating the channel ID.
|
|
var edgeKey [33 + 8]byte
|
|
copy(edgeKey[33:], chanID)
|
|
|
|
// With the latter half constructed, copy over the first public key to
|
|
// delete the edge in this direction, then the second to delete the
|
|
// edge in the opposite direction.
|
|
copy(edgeKey[:33], nodeKeys[:33])
|
|
if edges.Get(edgeKey[:]) != nil {
|
|
if err := edges.Delete(edgeKey[:]); err != nil {
|
|
return err
|
|
}
|
|
}
|
|
copy(edgeKey[:33], nodeKeys[33:])
|
|
if edges.Get(edgeKey[:]) != nil {
|
|
if err := edges.Delete(edgeKey[:]); err != nil {
|
|
return err
|
|
}
|
|
}
|
|
|
|
// Finally, with the edge data deleted, we can purge the information
|
|
// from the two edge indexes.
|
|
if err := edgeIndex.Delete(chanID); err != nil {
|
|
return err
|
|
}
|
|
return chanIndex.Delete(b.Bytes())
|
|
}
|
|
|
|
// UpdateEdgePolicy updates the edge routing policy for a single directed edge
|
|
// within the database for the referenced channel. The `flags` attribute within
|
|
// the ChannelEdgePolicy determines which of the directed edges are being
|
|
// updated. If the flag is 1, then the first node's information is being
|
|
// updated, otherwise it's the second node's information. The node ordering is
|
|
// determined by the lexicographical ordering of the identity public keys of
|
|
// the nodes on either side of the channel.
|
|
func (c *ChannelGraph) UpdateEdgePolicy(edge *ChannelEdgePolicy) error {
|
|
return c.db.Update(func(tx *bolt.Tx) error {
|
|
edges, err := tx.CreateBucketIfNotExists(edgeBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
edgeIndex, err := edges.CreateBucketIfNotExists(edgeIndexBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
// Create the channelID key be converting the channel ID
|
|
// integer into a byte slice.
|
|
var chanID [8]byte
|
|
byteOrder.PutUint64(chanID[:], edge.ChannelID)
|
|
|
|
// With the channel ID, we then fetch the value storing the two
|
|
// nodes which connect this channel edge.
|
|
nodeInfo := edgeIndex.Get(chanID[:])
|
|
if nodeInfo == nil {
|
|
return ErrEdgeNotFound
|
|
}
|
|
|
|
// Depending on the flags value passed above, either the first
|
|
// or second edge policy is being updated.
|
|
var fromNode, toNode []byte
|
|
if edge.Flags&lnwire.ChanUpdateDirection == 0 {
|
|
fromNode = nodeInfo[:33]
|
|
toNode = nodeInfo[33:67]
|
|
} else {
|
|
fromNode = nodeInfo[33:67]
|
|
toNode = nodeInfo[:33]
|
|
}
|
|
|
|
// Finally, with the direction of the edge being updated
|
|
// identified, we update the on-disk edge representation.
|
|
return putChanEdgePolicy(edges, edge, fromNode, toNode)
|
|
})
|
|
}
|
|
|
|
// LightningNode represents an individual vertex/node within the channel graph.
|
|
// A node is connected to other nodes by one or more channel edges emanating
|
|
// from it. As the graph is directed, a node will also have an incoming edge
|
|
// attached to it for each outgoing edge.
|
|
type LightningNode struct {
|
|
// PubKeyBytes is the raw bytes of the public key of the target node.
|
|
PubKeyBytes [33]byte
|
|
pubKey *btcec.PublicKey
|
|
|
|
// HaveNodeAnnouncement indicates whether we received a node
|
|
// announcement for this particular node. If true, the remaining fields
|
|
// will be set, if false only the PubKey is known for this node.
|
|
HaveNodeAnnouncement bool
|
|
|
|
// LastUpdate is the last time the vertex information for this node has
|
|
// been updated.
|
|
LastUpdate time.Time
|
|
|
|
// Address is the TCP address this node is reachable over.
|
|
Addresses []net.Addr
|
|
|
|
// Color is the selected color for the node.
|
|
Color color.RGBA
|
|
|
|
// Alias is a nick-name for the node. The alias can be used to confirm
|
|
// a node's identity or to serve as a short ID for an address book.
|
|
Alias string
|
|
|
|
// AuthSigBytes is the raw signature under the advertised public key
|
|
// which serves to authenticate the attributes announced by this node.
|
|
AuthSigBytes []byte
|
|
|
|
// Features is the list of protocol features supported by this node.
|
|
Features *lnwire.FeatureVector
|
|
|
|
db *DB
|
|
|
|
// TODO(roasbeef): discovery will need storage to keep it's last IP
|
|
// address and re-announce if interface changes?
|
|
|
|
// TODO(roasbeef): add update method and fetch?
|
|
}
|
|
|
|
// PubKey is the node's long-term identity public key. This key will be used to
|
|
// authenticated any advertisements/updates sent by the node.
|
|
//
|
|
// NOTE: By having this method to access an attribute, we ensure we only need
|
|
// to fully deserialize the pubkey if absolutely necessary.
|
|
func (l *LightningNode) PubKey() (*btcec.PublicKey, error) {
|
|
if l.pubKey != nil {
|
|
return l.pubKey, nil
|
|
}
|
|
|
|
key, err := btcec.ParsePubKey(l.PubKeyBytes[:], btcec.S256())
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
l.pubKey = key
|
|
l.pubKey.Curve = nil
|
|
|
|
return key, nil
|
|
}
|
|
|
|
// AuthSig is a signature under the advertised public key which serves to
|
|
// authenticate the attributes announced by this node.
|
|
//
|
|
// NOTE: By having this method to access an attribute, we ensure we only need
|
|
// to fully deserialize the signature if absolutely necessary.
|
|
func (l *LightningNode) AuthSig() (*btcec.Signature, error) {
|
|
return btcec.ParseSignature(l.AuthSigBytes, btcec.S256())
|
|
}
|
|
|
|
// AddPubKey is a setter-link method that can be used to swap out the public
|
|
// key for a node.
|
|
func (l *LightningNode) AddPubKey(key *btcec.PublicKey) {
|
|
l.pubKey = key
|
|
copy(l.PubKeyBytes[:], key.SerializeCompressed())
|
|
}
|
|
|
|
// FetchLightningNode attempts to look up a target node by its identity public
|
|
// key. If the node isn't found in the database, then ErrGraphNodeNotFound is
|
|
// returned.
|
|
func (c *ChannelGraph) FetchLightningNode(pub *btcec.PublicKey) (*LightningNode, error) {
|
|
var node *LightningNode
|
|
nodePub := pub.SerializeCompressed()
|
|
err := c.db.View(func(tx *bolt.Tx) error {
|
|
// First grab the nodes bucket which stores the mapping from
|
|
// pubKey to node information.
|
|
nodes := tx.Bucket(nodeBucket)
|
|
if nodes == nil {
|
|
return ErrGraphNotFound
|
|
}
|
|
|
|
// If a key for this serialized public key isn't found, then
|
|
// the target node doesn't exist within the database.
|
|
nodeBytes := nodes.Get(nodePub)
|
|
if nodeBytes == nil {
|
|
return ErrGraphNodeNotFound
|
|
}
|
|
|
|
// If the node is found, then we can de deserialize the node
|
|
// information to return to the user.
|
|
nodeReader := bytes.NewReader(nodeBytes)
|
|
n, err := deserializeLightningNode(nodeReader)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
n.db = c.db
|
|
|
|
node = &n
|
|
|
|
return nil
|
|
})
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
return node, nil
|
|
}
|
|
|
|
// HasLightningNode determines if the graph has a vertex identified by the
|
|
// target node identity public key. If the node exists in the database, a
|
|
// timestamp of when the data for the node was lasted updated is returned along
|
|
// with a true boolean. Otherwise, an empty time.Time is returned with a false
|
|
// boolean.
|
|
func (c *ChannelGraph) HasLightningNode(nodePub [33]byte) (time.Time, bool, error) {
|
|
var (
|
|
updateTime time.Time
|
|
exists bool
|
|
)
|
|
|
|
err := c.db.View(func(tx *bolt.Tx) error {
|
|
// First grab the nodes bucket which stores the mapping from
|
|
// pubKey to node information.
|
|
nodes := tx.Bucket(nodeBucket)
|
|
if nodes == nil {
|
|
return ErrGraphNotFound
|
|
}
|
|
|
|
// If a key for this serialized public key isn't found, we can
|
|
// exit early.
|
|
nodeBytes := nodes.Get(nodePub[:])
|
|
if nodeBytes == nil {
|
|
exists = false
|
|
return nil
|
|
}
|
|
|
|
// Otherwise we continue on to obtain the time stamp
|
|
// representing the last time the data for this node was
|
|
// updated.
|
|
nodeReader := bytes.NewReader(nodeBytes)
|
|
node, err := deserializeLightningNode(nodeReader)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
exists = true
|
|
updateTime = node.LastUpdate
|
|
return nil
|
|
})
|
|
if err != nil {
|
|
return time.Time{}, exists, nil
|
|
}
|
|
|
|
return updateTime, exists, nil
|
|
}
|
|
|
|
// ForEachChannel iterates through all channels of this node, executing the
|
|
// passed callback with an edge info structure and the policies of each end
|
|
// of the channel. The first edge policy is the outgoing edge *to* the
|
|
// the connecting node, while the second is the incoming edge *from* the
|
|
// connecting node. If the callback returns an error, then the iteration is
|
|
// halted with the error propagated back up to the caller.
|
|
//
|
|
// Unknown policies are passed into the callback as nil values.
|
|
//
|
|
// If the caller wishes to re-use an existing boltdb transaction, then it
|
|
// should be passed as the first argument. Otherwise the first argument should
|
|
// be nil and a fresh transaction will be created to execute the graph
|
|
// traversal.
|
|
func (l *LightningNode) ForEachChannel(tx *bolt.Tx,
|
|
cb func(*bolt.Tx, *ChannelEdgeInfo, *ChannelEdgePolicy, *ChannelEdgePolicy) error) error {
|
|
|
|
nodePub := l.PubKeyBytes[:]
|
|
|
|
traversal := func(tx *bolt.Tx) error {
|
|
nodes := tx.Bucket(nodeBucket)
|
|
if nodes == nil {
|
|
return ErrGraphNotFound
|
|
}
|
|
edges := tx.Bucket(edgeBucket)
|
|
if edges == nil {
|
|
return ErrGraphNotFound
|
|
}
|
|
edgeIndex := edges.Bucket(edgeIndexBucket)
|
|
if edgeIndex == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
|
|
// In order to reach all the edges for this node, we take
|
|
// advantage of the construction of the key-space within the
|
|
// edge bucket. The keys are stored in the form: pubKey ||
|
|
// chanID. Therefore, starting from a chanID of zero, we can
|
|
// scan forward in the bucket, grabbing all the edges for the
|
|
// node. Once the prefix no longer matches, then we know we're
|
|
// done.
|
|
var nodeStart [33 + 8]byte
|
|
copy(nodeStart[:], nodePub)
|
|
copy(nodeStart[33:], chanStart[:])
|
|
|
|
// Starting from the key pubKey || 0, we seek forward in the
|
|
// bucket until the retrieved key no longer has the public key
|
|
// as its prefix. This indicates that we've stepped over into
|
|
// another node's edges, so we can terminate our scan.
|
|
edgeCursor := edges.Cursor()
|
|
for nodeEdge, _ := edgeCursor.Seek(nodeStart[:]); bytes.HasPrefix(nodeEdge, nodePub); nodeEdge, _ = edgeCursor.Next() {
|
|
// If the prefix still matches, the channel id is
|
|
// returned in nodeEdge. Channel id is used to lookup
|
|
// the node at the other end of the channel and both
|
|
// edge policies.
|
|
chanID := nodeEdge[33:]
|
|
edgeInfo, err := fetchChanEdgeInfo(edgeIndex, chanID)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
outgoingPolicy, err := fetchChanEdgePolicy(
|
|
edges, chanID, nodePub, nodes,
|
|
)
|
|
|
|
otherNode, err := edgeInfo.OtherNodeKeyBytes(nodePub)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
incomingPolicy, err := fetchChanEdgePolicy(
|
|
edges, chanID, otherNode, nodes,
|
|
)
|
|
|
|
// Finally, we execute the callback.
|
|
err = cb(tx, &edgeInfo, outgoingPolicy, incomingPolicy)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
}
|
|
|
|
return nil
|
|
}
|
|
|
|
// If no transaction was provided, then we'll create a new transaction
|
|
// to execute the transaction within.
|
|
if tx == nil {
|
|
return l.db.View(traversal)
|
|
}
|
|
|
|
// Otherwise, we re-use the existing transaction to execute the graph
|
|
// traversal.
|
|
return traversal(tx)
|
|
}
|
|
|
|
// ChannelEdgeInfo represents a fully authenticated channel along with all its
|
|
// unique attributes. Once an authenticated channel announcement has been
|
|
// processed on the network, then an instance of ChannelEdgeInfo encapsulating
|
|
// the channels attributes is stored. The other portions relevant to routing
|
|
// policy of a channel are stored within a ChannelEdgePolicy for each direction
|
|
// of the channel.
|
|
type ChannelEdgeInfo struct {
|
|
// ChannelID is the unique channel ID for the channel. The first 3
|
|
// bytes are the block height, the next 3 the index within the block,
|
|
// and the last 2 bytes are the output index for the channel.
|
|
ChannelID uint64
|
|
|
|
// ChainHash is the hash that uniquely identifies the chain that this
|
|
// channel was opened within.
|
|
//
|
|
// TODO(roasbeef): need to modify db keying for multi-chain
|
|
// * must add chain hash to prefix as well
|
|
ChainHash chainhash.Hash
|
|
|
|
// NodeKey1Bytes is the raw public key of the first node.
|
|
NodeKey1Bytes [33]byte
|
|
nodeKey1 *btcec.PublicKey
|
|
|
|
// NodeKey2Bytes is the raw public key of the first node.
|
|
NodeKey2Bytes [33]byte
|
|
nodeKey2 *btcec.PublicKey
|
|
|
|
// BitcoinKey1Bytes is the raw public key of the first node.
|
|
BitcoinKey1Bytes [33]byte
|
|
bitcoinKey1 *btcec.PublicKey
|
|
|
|
// BitcoinKey2Bytes is the raw public key of the first node.
|
|
BitcoinKey2Bytes [33]byte
|
|
bitcoinKey2 *btcec.PublicKey
|
|
|
|
// Features is an opaque byte slice that encodes the set of channel
|
|
// specific features that this channel edge supports.
|
|
Features []byte
|
|
|
|
// AuthProof is the authentication proof for this channel. This proof
|
|
// contains a set of signatures binding four identities, which attests
|
|
// to the legitimacy of the advertised channel.
|
|
AuthProof *ChannelAuthProof
|
|
|
|
// ChannelPoint is the funding outpoint of the channel. This can be
|
|
// used to uniquely identify the channel within the channel graph.
|
|
ChannelPoint wire.OutPoint
|
|
|
|
// Capacity is the total capacity of the channel, this is determined by
|
|
// the value output in the outpoint that created this channel.
|
|
Capacity btcutil.Amount
|
|
}
|
|
|
|
// AddNodeKeys is a setter-like method that can be used to replace the set of
|
|
// keys for the target ChannelEdgeInfo.
|
|
func (c *ChannelEdgeInfo) AddNodeKeys(nodeKey1, nodeKey2, bitcoinKey1,
|
|
bitcoinKey2 *btcec.PublicKey) {
|
|
|
|
c.nodeKey1 = nodeKey1
|
|
copy(c.NodeKey1Bytes[:], c.nodeKey1.SerializeCompressed())
|
|
|
|
c.nodeKey2 = nodeKey2
|
|
copy(c.NodeKey2Bytes[:], nodeKey2.SerializeCompressed())
|
|
|
|
c.bitcoinKey1 = bitcoinKey1
|
|
copy(c.BitcoinKey1Bytes[:], c.bitcoinKey1.SerializeCompressed())
|
|
|
|
c.bitcoinKey2 = bitcoinKey2
|
|
copy(c.BitcoinKey2Bytes[:], bitcoinKey2.SerializeCompressed())
|
|
}
|
|
|
|
// NodeKey1 is the identity public key of the "first" node that was involved in
|
|
// the creation of this channel. A node is considered "first" if the
|
|
// lexicographical ordering the its serialized public key is "smaller" than
|
|
// that of the other node involved in channel creation.
|
|
//
|
|
// NOTE: By having this method to access an attribute, we ensure we only need
|
|
// to fully deserialize the pubkey if absolutely necessary.
|
|
func (c *ChannelEdgeInfo) NodeKey1() (*btcec.PublicKey, error) {
|
|
if c.nodeKey1 != nil {
|
|
return c.nodeKey1, nil
|
|
}
|
|
|
|
key, err := btcec.ParsePubKey(c.NodeKey1Bytes[:], btcec.S256())
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
c.nodeKey1 = key
|
|
|
|
return key, nil
|
|
}
|
|
|
|
// NodeKey2 is the identity public key of the "second" node that was
|
|
// involved in the creation of this channel. A node is considered
|
|
// "second" if the lexicographical ordering the its serialized public
|
|
// key is "larger" than that of the other node involved in channel
|
|
// creation.
|
|
//
|
|
// NOTE: By having this method to access an attribute, we ensure we only need
|
|
// to fully deserialize the pubkey if absolutely necessary.
|
|
func (c *ChannelEdgeInfo) NodeKey2() (*btcec.PublicKey, error) {
|
|
if c.nodeKey2 != nil {
|
|
return c.nodeKey2, nil
|
|
}
|
|
|
|
key, err := btcec.ParsePubKey(c.NodeKey2Bytes[:], btcec.S256())
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
c.nodeKey2 = key
|
|
|
|
return key, nil
|
|
}
|
|
|
|
// BitcoinKey1 is the Bitcoin multi-sig key belonging to the first
|
|
// node, that was involved in the funding transaction that originally
|
|
// created the channel that this struct represents.
|
|
//
|
|
// NOTE: By having this method to access an attribute, we ensure we only need
|
|
// to fully deserialize the pubkey if absolutely necessary.
|
|
func (c *ChannelEdgeInfo) BitcoinKey1() (*btcec.PublicKey, error) {
|
|
if c.bitcoinKey1 != nil {
|
|
return c.bitcoinKey1, nil
|
|
}
|
|
|
|
key, err := btcec.ParsePubKey(c.BitcoinKey1Bytes[:], btcec.S256())
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
c.bitcoinKey1 = key
|
|
|
|
return key, nil
|
|
}
|
|
|
|
// BitcoinKey2 is the Bitcoin multi-sig key belonging to the second
|
|
// node, that was involved in the funding transaction that originally
|
|
// created the channel that this struct represents.
|
|
//
|
|
// NOTE: By having this method to access an attribute, we ensure we only need
|
|
// to fully deserialize the pubkey if absolutely necessary.
|
|
func (c *ChannelEdgeInfo) BitcoinKey2() (*btcec.PublicKey, error) {
|
|
if c.bitcoinKey2 != nil {
|
|
return c.bitcoinKey2, nil
|
|
}
|
|
|
|
key, err := btcec.ParsePubKey(c.BitcoinKey2Bytes[:], btcec.S256())
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
c.bitcoinKey2 = key
|
|
|
|
return key, nil
|
|
}
|
|
|
|
// OtherNodeKeyBytes returns the node key bytes of the other end of
|
|
// the channel.
|
|
func (c *ChannelEdgeInfo) OtherNodeKeyBytes(thisNodeKey []byte) (
|
|
[]byte, error) {
|
|
|
|
switch {
|
|
case bytes.Equal(c.NodeKey1Bytes[:], thisNodeKey):
|
|
return c.NodeKey2Bytes[:], nil
|
|
case bytes.Equal(c.NodeKey2Bytes[:], thisNodeKey):
|
|
return c.NodeKey1Bytes[:], nil
|
|
default:
|
|
return nil, fmt.Errorf("Node not participating in this channel")
|
|
}
|
|
}
|
|
|
|
// ChannelAuthProof is the authentication proof (the signature portion) for a
|
|
// channel. Using the four signatures contained in the struct, and some
|
|
// auxiliary knowledge (the funding script, node identities, and outpoint) nodes
|
|
// on the network are able to validate the authenticity and existence of a
|
|
// channel. Each of these signatures signs the following digest: chanID ||
|
|
// nodeID1 || nodeID2 || bitcoinKey1|| bitcoinKey2 || 2-byte-feature-len ||
|
|
// features.
|
|
type ChannelAuthProof struct {
|
|
// nodeSig1 is a cached instance of the first node signature.
|
|
nodeSig1 *btcec.Signature
|
|
|
|
// NodeSig1Bytes are the raw bytes of the first node signature encoded
|
|
// in DER format.
|
|
NodeSig1Bytes []byte
|
|
|
|
// nodeSig2 is a cached instance of the second node signature.
|
|
nodeSig2 *btcec.Signature
|
|
|
|
// NodeSig2Bytes are the raw bytes of the second node signature
|
|
// encoded in DER format.
|
|
NodeSig2Bytes []byte
|
|
|
|
// bitcoinSig1 is a cached instance of the first bitcoin signature.
|
|
bitcoinSig1 *btcec.Signature
|
|
|
|
// BitcoinSig1Bytes are the raw bytes of the first bitcoin signature
|
|
// encoded in DER format.
|
|
BitcoinSig1Bytes []byte
|
|
|
|
// bitcoinSig2 is a cached instance of the second bitcoin signature.
|
|
bitcoinSig2 *btcec.Signature
|
|
|
|
// BitcoinSig2Bytes are the raw bytes of the second bitcoin signature
|
|
// encoded in DER format.
|
|
BitcoinSig2Bytes []byte
|
|
}
|
|
|
|
// Node1Sig is the signature using the identity key of the node that is first
|
|
// in a lexicographical ordering of the serialized public keys of the two nodes
|
|
// that created the channel.
|
|
//
|
|
// NOTE: By having this method to access an attribute, we ensure we only need
|
|
// to fully deserialize the signature if absolutely necessary.
|
|
func (c *ChannelAuthProof) Node1Sig() (*btcec.Signature, error) {
|
|
if c.nodeSig1 != nil {
|
|
return c.nodeSig1, nil
|
|
}
|
|
|
|
sig, err := btcec.ParseSignature(c.NodeSig1Bytes, btcec.S256())
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
c.nodeSig1 = sig
|
|
|
|
return sig, nil
|
|
}
|
|
|
|
// Node2Sig is the signature using the identity key of the node that is second
|
|
// in a lexicographical ordering of the serialized public keys of the two nodes
|
|
// that created the channel.
|
|
//
|
|
// NOTE: By having this method to access an attribute, we ensure we only need
|
|
// to fully deserialize the signature if absolutely necessary.
|
|
func (c *ChannelAuthProof) Node2Sig() (*btcec.Signature, error) {
|
|
if c.nodeSig2 != nil {
|
|
return c.nodeSig2, nil
|
|
}
|
|
|
|
sig, err := btcec.ParseSignature(c.NodeSig2Bytes, btcec.S256())
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
c.nodeSig2 = sig
|
|
|
|
return sig, nil
|
|
}
|
|
|
|
// BitcoinSig1 is the signature using the public key of the first node that was
|
|
// used in the channel's multi-sig output.
|
|
//
|
|
// NOTE: By having this method to access an attribute, we ensure we only need
|
|
// to fully deserialize the signature if absolutely necessary.
|
|
func (c *ChannelAuthProof) BitcoinSig1() (*btcec.Signature, error) {
|
|
if c.bitcoinSig1 != nil {
|
|
return c.bitcoinSig1, nil
|
|
}
|
|
|
|
sig, err := btcec.ParseSignature(c.BitcoinSig1Bytes, btcec.S256())
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
c.bitcoinSig1 = sig
|
|
|
|
return sig, nil
|
|
}
|
|
|
|
// BitcoinSig2 is the signature using the public key of the second node that
|
|
// was used in the channel's multi-sig output.
|
|
//
|
|
// NOTE: By having this method to access an attribute, we ensure we only need
|
|
// to fully deserialize the signature if absolutely necessary.
|
|
func (c *ChannelAuthProof) BitcoinSig2() (*btcec.Signature, error) {
|
|
if c.bitcoinSig2 != nil {
|
|
return c.bitcoinSig2, nil
|
|
}
|
|
|
|
sig, err := btcec.ParseSignature(c.BitcoinSig2Bytes, btcec.S256())
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
c.bitcoinSig2 = sig
|
|
|
|
return sig, nil
|
|
}
|
|
|
|
// IsEmpty check is the authentication proof is empty Proof is empty if at
|
|
// least one of the signatures are equal to nil.
|
|
func (c *ChannelAuthProof) IsEmpty() bool {
|
|
return len(c.NodeSig1Bytes) == 0 ||
|
|
len(c.NodeSig2Bytes) == 0 ||
|
|
len(c.BitcoinSig1Bytes) == 0 ||
|
|
len(c.BitcoinSig2Bytes) == 0
|
|
}
|
|
|
|
// ChannelEdgePolicy represents a *directed* edge within the channel graph. For
|
|
// each channel in the database, there are two distinct edges: one for each
|
|
// possible direction of travel along the channel. The edges themselves hold
|
|
// information concerning fees, and minimum time-lock information which is
|
|
// utilized during path finding.
|
|
type ChannelEdgePolicy struct {
|
|
// SigBytes is the raw bytes of the signature of the channel edge
|
|
// policy. We'll only parse these if the caller needs to access the
|
|
// signature for validation purposes.
|
|
SigBytes []byte
|
|
|
|
// sig is a cached fully parsed signature.
|
|
sig *btcec.Signature
|
|
|
|
// ChannelID is the unique channel ID for the channel. The first 3
|
|
// bytes are the block height, the next 3 the index within the block,
|
|
// and the last 2 bytes are the output index for the channel.
|
|
ChannelID uint64
|
|
|
|
// LastUpdate is the last time an authenticated edge for this channel
|
|
// was received.
|
|
LastUpdate time.Time
|
|
|
|
// Flags is a bitfield which signals the capabilities of the channel as
|
|
// well as the directed edge this update applies to.
|
|
Flags lnwire.ChanUpdateFlag
|
|
|
|
// TimeLockDelta is the number of blocks this node will subtract from
|
|
// the expiry of an incoming HTLC. This value expresses the time buffer
|
|
// the node would like to HTLC exchanges.
|
|
TimeLockDelta uint16
|
|
|
|
// MinHTLC is the smallest value HTLC this node will accept, expressed
|
|
// in millisatoshi.
|
|
MinHTLC lnwire.MilliSatoshi
|
|
|
|
// FeeBaseMSat is the base HTLC fee that will be charged for forwarding
|
|
// ANY HTLC, expressed in mSAT's.
|
|
FeeBaseMSat lnwire.MilliSatoshi
|
|
|
|
// FeeProportionalMillionths is the rate that the node will charge for
|
|
// HTLCs for each millionth of a satoshi forwarded.
|
|
FeeProportionalMillionths lnwire.MilliSatoshi
|
|
|
|
// Node is the LightningNode that this directed edge leads to. Using
|
|
// this pointer the channel graph can further be traversed.
|
|
Node *LightningNode
|
|
|
|
db *DB
|
|
}
|
|
|
|
// Signature is a channel announcement signature, which is needed for proper
|
|
// edge policy announcement.
|
|
//
|
|
// NOTE: By having this method to access an attribute, we ensure we only need
|
|
// to fully deserialize the signature if absolutely necessary.
|
|
func (c *ChannelEdgePolicy) Signature() (*btcec.Signature, error) {
|
|
if c.sig != nil {
|
|
return c.sig, nil
|
|
}
|
|
|
|
sig, err := btcec.ParseSignature(c.SigBytes, btcec.S256())
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
c.sig = sig
|
|
|
|
return sig, nil
|
|
}
|
|
|
|
// FetchChannelEdgesByOutpoint attempts to lookup the two directed edges for
|
|
// the channel identified by the funding outpoint. If the channel can't be
|
|
// found, then ErrEdgeNotFound is returned. A struct which houses the general
|
|
// information for the channel itself is returned as well as two structs that
|
|
// contain the routing policies for the channel in either direction.
|
|
func (c *ChannelGraph) FetchChannelEdgesByOutpoint(op *wire.OutPoint) (*ChannelEdgeInfo, *ChannelEdgePolicy, *ChannelEdgePolicy, error) {
|
|
|
|
var (
|
|
edgeInfo *ChannelEdgeInfo
|
|
policy1 *ChannelEdgePolicy
|
|
policy2 *ChannelEdgePolicy
|
|
)
|
|
|
|
err := c.db.View(func(tx *bolt.Tx) error {
|
|
// First, grab the node bucket. This will be used to populate
|
|
// the Node pointers in each edge read from disk.
|
|
nodes := tx.Bucket(nodeBucket)
|
|
if nodes == nil {
|
|
return ErrGraphNotFound
|
|
}
|
|
|
|
// Next, grab the edge bucket which stores the edges, and also
|
|
// the index itself so we can group the directed edges together
|
|
// logically.
|
|
edges := tx.Bucket(edgeBucket)
|
|
if edges == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
edgeIndex := edges.Bucket(edgeIndexBucket)
|
|
if edgeIndex == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
|
|
// If the channel's outpoint doesn't exist within the outpoint
|
|
// index, then the edge does not exist.
|
|
chanIndex := edges.Bucket(channelPointBucket)
|
|
if chanIndex == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
var b bytes.Buffer
|
|
if err := writeOutpoint(&b, op); err != nil {
|
|
return err
|
|
}
|
|
chanID := chanIndex.Get(b.Bytes())
|
|
if chanID == nil {
|
|
return ErrEdgeNotFound
|
|
}
|
|
|
|
// If the channel is found to exists, then we'll first retrieve
|
|
// the general information for the channel.
|
|
edge, err := fetchChanEdgeInfo(edgeIndex, chanID)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
edgeInfo = &edge
|
|
|
|
// Once we have the information about the channels' parameters,
|
|
// we'll fetch the routing policies for each for the directed
|
|
// edges.
|
|
e1, e2, err := fetchChanEdgePolicies(
|
|
edgeIndex, edges, nodes, chanID, c.db,
|
|
)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
policy1 = e1
|
|
policy2 = e2
|
|
return nil
|
|
})
|
|
if err != nil {
|
|
return nil, nil, nil, err
|
|
}
|
|
|
|
return edgeInfo, policy1, policy2, nil
|
|
}
|
|
|
|
// FetchChannelEdgesByID attempts to lookup the two directed edges for the
|
|
// channel identified by the channel ID. If the channel can't be found, then
|
|
// ErrEdgeNotFound is returned. A struct which houses the general information
|
|
// for the channel itself is returned as well as two structs that contain the
|
|
// routing policies for the channel in either direction.
|
|
func (c *ChannelGraph) FetchChannelEdgesByID(chanID uint64) (*ChannelEdgeInfo, *ChannelEdgePolicy, *ChannelEdgePolicy, error) {
|
|
|
|
var (
|
|
edgeInfo *ChannelEdgeInfo
|
|
policy1 *ChannelEdgePolicy
|
|
policy2 *ChannelEdgePolicy
|
|
channelID [8]byte
|
|
)
|
|
|
|
err := c.db.View(func(tx *bolt.Tx) error {
|
|
// First, grab the node bucket. This will be used to populate
|
|
// the Node pointers in each edge read from disk.
|
|
nodes := tx.Bucket(nodeBucket)
|
|
if nodes == nil {
|
|
return ErrGraphNotFound
|
|
}
|
|
|
|
// Next, grab the edge bucket which stores the edges, and also
|
|
// the index itself so we can group the directed edges together
|
|
// logically.
|
|
edges := tx.Bucket(edgeBucket)
|
|
if edges == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
edgeIndex := edges.Bucket(edgeIndexBucket)
|
|
if edgeIndex == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
|
|
byteOrder.PutUint64(channelID[:], chanID)
|
|
|
|
edge, err := fetchChanEdgeInfo(edgeIndex, channelID[:])
|
|
if err != nil {
|
|
return err
|
|
}
|
|
edgeInfo = &edge
|
|
|
|
e1, e2, err := fetchChanEdgePolicies(
|
|
edgeIndex, edges, nodes, channelID[:], c.db,
|
|
)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
policy1 = e1
|
|
policy2 = e2
|
|
return nil
|
|
})
|
|
if err != nil {
|
|
return nil, nil, nil, err
|
|
}
|
|
|
|
return edgeInfo, policy1, policy2, nil
|
|
}
|
|
|
|
// genMultiSigP2WSH generates the p2wsh'd multisig script for 2 of 2 pubkeys.
|
|
func genMultiSigP2WSH(aPub, bPub []byte) ([]byte, error) {
|
|
if len(aPub) != 33 || len(bPub) != 33 {
|
|
return nil, fmt.Errorf("Pubkey size error. Compressed " +
|
|
"pubkeys only")
|
|
}
|
|
|
|
// Swap to sort pubkeys if needed. Keys are sorted in lexicographical
|
|
// order. The signatures within the scriptSig must also adhere to the
|
|
// order, ensuring that the signatures for each public key appears in
|
|
// the proper order on the stack.
|
|
if bytes.Compare(aPub, bPub) == 1 {
|
|
aPub, bPub = bPub, aPub
|
|
}
|
|
|
|
// First, we'll generate the witness script for the multi-sig.
|
|
bldr := txscript.NewScriptBuilder()
|
|
bldr.AddOp(txscript.OP_2)
|
|
bldr.AddData(aPub) // Add both pubkeys (sorted).
|
|
bldr.AddData(bPub)
|
|
bldr.AddOp(txscript.OP_2)
|
|
bldr.AddOp(txscript.OP_CHECKMULTISIG)
|
|
witnessScript, err := bldr.Script()
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
// With the witness script generated, we'll now turn it into a p2sh
|
|
// script:
|
|
// * OP_0 <sha256(script)>
|
|
bldr = txscript.NewScriptBuilder()
|
|
bldr.AddOp(txscript.OP_0)
|
|
scriptHash := sha256.Sum256(witnessScript)
|
|
bldr.AddData(scriptHash[:])
|
|
|
|
return bldr.Script()
|
|
}
|
|
|
|
// EdgePoint couples the outpoint of a channel with the funding script that it
|
|
// creates. The FilteredChainView will use this to watch for spends of this
|
|
// edge point on chain. We require both of these values as depending on the
|
|
// concrete implementation, either the pkScript, or the out point will be used.
|
|
type EdgePoint struct {
|
|
// FundingPkScript is the p2wsh multi-sig script of the target channel.
|
|
FundingPkScript []byte
|
|
|
|
// OutPoint is the outpoint of the target channel.
|
|
OutPoint wire.OutPoint
|
|
}
|
|
|
|
// ChannelView returns the verifiable edge information for each active channel
|
|
// within the known channel graph. The set of UTXO's (along with their scripts)
|
|
// returned are the ones that need to be watched on chain to detect channel
|
|
// closes on the resident blockchain.
|
|
func (c *ChannelGraph) ChannelView() ([]EdgePoint, error) {
|
|
var edgePoints []EdgePoint
|
|
if err := c.db.View(func(tx *bolt.Tx) error {
|
|
// We're going to iterate over the entire channel index, so
|
|
// we'll need to fetch the edgeBucket to get to the index as
|
|
// it's a sub-bucket.
|
|
edges := tx.Bucket(edgeBucket)
|
|
if edges == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
chanIndex := edges.Bucket(channelPointBucket)
|
|
if chanIndex == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
edgeIndex := edges.Bucket(edgeIndexBucket)
|
|
if edgeIndex == nil {
|
|
return ErrGraphNoEdgesFound
|
|
}
|
|
|
|
// Once we have the proper bucket, we'll range over each key
|
|
// (which is the channel point for the channel) and decode it,
|
|
// accumulating each entry.
|
|
return chanIndex.ForEach(func(chanPointBytes, chanID []byte) error {
|
|
chanPointReader := bytes.NewReader(chanPointBytes)
|
|
|
|
var chanPoint wire.OutPoint
|
|
err := readOutpoint(chanPointReader, &chanPoint)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
edgeInfo, err := fetchChanEdgeInfo(
|
|
edgeIndex, chanID,
|
|
)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
pkScript, err := genMultiSigP2WSH(
|
|
edgeInfo.BitcoinKey1Bytes[:],
|
|
edgeInfo.BitcoinKey2Bytes[:],
|
|
)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
edgePoints = append(edgePoints, EdgePoint{
|
|
FundingPkScript: pkScript,
|
|
OutPoint: chanPoint,
|
|
})
|
|
|
|
return nil
|
|
})
|
|
}); err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
return edgePoints, nil
|
|
}
|
|
|
|
// NewChannelEdgePolicy returns a new blank ChannelEdgePolicy.
|
|
func (c *ChannelGraph) NewChannelEdgePolicy() *ChannelEdgePolicy {
|
|
return &ChannelEdgePolicy{db: c.db}
|
|
}
|
|
|
|
func putLightningNode(nodeBucket *bolt.Bucket, aliasBucket *bolt.Bucket,
|
|
updateIndex *bolt.Bucket, node *LightningNode) error {
|
|
|
|
var (
|
|
scratch [16]byte
|
|
b bytes.Buffer
|
|
)
|
|
|
|
pub, err := node.PubKey()
|
|
if err != nil {
|
|
return err
|
|
}
|
|
nodePub := pub.SerializeCompressed()
|
|
|
|
// If the node has the update time set, write it, else write 0.
|
|
updateUnix := uint64(0)
|
|
if node.LastUpdate.Unix() > 0 {
|
|
updateUnix = uint64(node.LastUpdate.Unix())
|
|
}
|
|
|
|
byteOrder.PutUint64(scratch[:8], updateUnix)
|
|
if _, err := b.Write(scratch[:8]); err != nil {
|
|
return err
|
|
}
|
|
|
|
if _, err := b.Write(nodePub); err != nil {
|
|
return err
|
|
}
|
|
|
|
// If we got a node announcement for this node, we will have the rest
|
|
// of the data available. If not we don't have more data to write.
|
|
if !node.HaveNodeAnnouncement {
|
|
// Write HaveNodeAnnouncement=0.
|
|
byteOrder.PutUint16(scratch[:2], 0)
|
|
if _, err := b.Write(scratch[:2]); err != nil {
|
|
return err
|
|
}
|
|
|
|
return nodeBucket.Put(nodePub, b.Bytes())
|
|
}
|
|
|
|
// Write HaveNodeAnnouncement=1.
|
|
byteOrder.PutUint16(scratch[:2], 1)
|
|
if _, err := b.Write(scratch[:2]); err != nil {
|
|
return err
|
|
}
|
|
|
|
if err := binary.Write(&b, byteOrder, node.Color.R); err != nil {
|
|
return err
|
|
}
|
|
if err := binary.Write(&b, byteOrder, node.Color.G); err != nil {
|
|
return err
|
|
}
|
|
if err := binary.Write(&b, byteOrder, node.Color.B); err != nil {
|
|
return err
|
|
}
|
|
|
|
if err := wire.WriteVarString(&b, 0, node.Alias); err != nil {
|
|
return err
|
|
}
|
|
|
|
if err := node.Features.Encode(&b); err != nil {
|
|
return err
|
|
}
|
|
|
|
numAddresses := uint16(len(node.Addresses))
|
|
byteOrder.PutUint16(scratch[:2], numAddresses)
|
|
if _, err := b.Write(scratch[:2]); err != nil {
|
|
return err
|
|
}
|
|
|
|
for _, address := range node.Addresses {
|
|
if err := serializeAddr(&b, address); err != nil {
|
|
return err
|
|
}
|
|
}
|
|
|
|
err = wire.WriteVarBytes(&b, 0, node.AuthSigBytes)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
if err := aliasBucket.Put(nodePub, []byte(node.Alias)); err != nil {
|
|
return err
|
|
}
|
|
|
|
// With the alias bucket updated, we'll now update the index that
|
|
// tracks the time series of node updates.
|
|
var indexKey [8 + 33]byte
|
|
byteOrder.PutUint64(indexKey[:8], updateUnix)
|
|
copy(indexKey[8:], nodePub)
|
|
|
|
// If there was already an old index entry for this node, then we'll
|
|
// delete the old one before we write the new entry.
|
|
if nodeBytes := nodeBucket.Get(nodePub); nodeBytes != nil {
|
|
// Extract out the old update time to we can reconstruct the
|
|
// prior index key to delete it from the index.
|
|
oldUpdateTime := nodeBytes[:8]
|
|
|
|
var oldIndexKey [8 + 33]byte
|
|
copy(oldIndexKey[:8], oldUpdateTime)
|
|
copy(oldIndexKey[8:], nodePub)
|
|
|
|
if err := updateIndex.Delete(oldIndexKey[:]); err != nil {
|
|
return err
|
|
}
|
|
}
|
|
|
|
if err := updateIndex.Put(indexKey[:], nil); err != nil {
|
|
return err
|
|
}
|
|
|
|
return nodeBucket.Put(nodePub, b.Bytes())
|
|
}
|
|
|
|
func fetchLightningNode(nodeBucket *bolt.Bucket,
|
|
nodePub []byte) (LightningNode, error) {
|
|
|
|
nodeBytes := nodeBucket.Get(nodePub)
|
|
if nodeBytes == nil {
|
|
return LightningNode{}, ErrGraphNodeNotFound
|
|
}
|
|
|
|
nodeReader := bytes.NewReader(nodeBytes)
|
|
return deserializeLightningNode(nodeReader)
|
|
}
|
|
|
|
func deserializeLightningNode(r io.Reader) (LightningNode, error) {
|
|
var (
|
|
node LightningNode
|
|
scratch [8]byte
|
|
err error
|
|
)
|
|
|
|
if _, err := r.Read(scratch[:]); err != nil {
|
|
return LightningNode{}, err
|
|
}
|
|
|
|
unix := int64(byteOrder.Uint64(scratch[:]))
|
|
node.LastUpdate = time.Unix(unix, 0)
|
|
|
|
if _, err := io.ReadFull(r, node.PubKeyBytes[:]); err != nil {
|
|
return LightningNode{}, err
|
|
}
|
|
|
|
if _, err := r.Read(scratch[:2]); err != nil {
|
|
return LightningNode{}, err
|
|
}
|
|
|
|
hasNodeAnn := byteOrder.Uint16(scratch[:2])
|
|
if hasNodeAnn == 1 {
|
|
node.HaveNodeAnnouncement = true
|
|
} else {
|
|
node.HaveNodeAnnouncement = false
|
|
}
|
|
|
|
// The rest of the data is optional, and will only be there if we got a node
|
|
// announcement for this node.
|
|
if !node.HaveNodeAnnouncement {
|
|
return node, nil
|
|
}
|
|
|
|
// We did get a node announcement for this node, so we'll have the rest
|
|
// of the data available.
|
|
if err := binary.Read(r, byteOrder, &node.Color.R); err != nil {
|
|
return LightningNode{}, err
|
|
}
|
|
if err := binary.Read(r, byteOrder, &node.Color.G); err != nil {
|
|
return LightningNode{}, err
|
|
}
|
|
if err := binary.Read(r, byteOrder, &node.Color.B); err != nil {
|
|
return LightningNode{}, err
|
|
}
|
|
|
|
node.Alias, err = wire.ReadVarString(r, 0)
|
|
if err != nil {
|
|
return LightningNode{}, err
|
|
}
|
|
|
|
fv := lnwire.NewFeatureVector(nil, lnwire.GlobalFeatures)
|
|
err = fv.Decode(r)
|
|
if err != nil {
|
|
return LightningNode{}, err
|
|
}
|
|
node.Features = fv
|
|
|
|
if _, err := r.Read(scratch[:2]); err != nil {
|
|
return LightningNode{}, err
|
|
}
|
|
numAddresses := int(byteOrder.Uint16(scratch[:2]))
|
|
|
|
var addresses []net.Addr
|
|
for i := 0; i < numAddresses; i++ {
|
|
address, err := deserializeAddr(r)
|
|
if err != nil {
|
|
return LightningNode{}, err
|
|
}
|
|
addresses = append(addresses, address)
|
|
}
|
|
node.Addresses = addresses
|
|
|
|
node.AuthSigBytes, err = wire.ReadVarBytes(r, 0, 80, "sig")
|
|
if err != nil {
|
|
return LightningNode{}, err
|
|
}
|
|
|
|
return node, nil
|
|
}
|
|
|
|
func putChanEdgeInfo(edgeIndex *bolt.Bucket, edgeInfo *ChannelEdgeInfo, chanID [8]byte) error {
|
|
var b bytes.Buffer
|
|
|
|
if _, err := b.Write(edgeInfo.NodeKey1Bytes[:]); err != nil {
|
|
return err
|
|
}
|
|
if _, err := b.Write(edgeInfo.NodeKey2Bytes[:]); err != nil {
|
|
return err
|
|
}
|
|
if _, err := b.Write(edgeInfo.BitcoinKey1Bytes[:]); err != nil {
|
|
return err
|
|
}
|
|
if _, err := b.Write(edgeInfo.BitcoinKey2Bytes[:]); err != nil {
|
|
return err
|
|
}
|
|
|
|
if err := wire.WriteVarBytes(&b, 0, edgeInfo.Features); err != nil {
|
|
return err
|
|
}
|
|
|
|
authProof := edgeInfo.AuthProof
|
|
var nodeSig1, nodeSig2, bitcoinSig1, bitcoinSig2 []byte
|
|
if authProof != nil {
|
|
nodeSig1 = authProof.NodeSig1Bytes
|
|
nodeSig2 = authProof.NodeSig2Bytes
|
|
bitcoinSig1 = authProof.BitcoinSig1Bytes
|
|
bitcoinSig2 = authProof.BitcoinSig2Bytes
|
|
}
|
|
|
|
if err := wire.WriteVarBytes(&b, 0, nodeSig1); err != nil {
|
|
return err
|
|
}
|
|
if err := wire.WriteVarBytes(&b, 0, nodeSig2); err != nil {
|
|
return err
|
|
}
|
|
if err := wire.WriteVarBytes(&b, 0, bitcoinSig1); err != nil {
|
|
return err
|
|
}
|
|
if err := wire.WriteVarBytes(&b, 0, bitcoinSig2); err != nil {
|
|
return err
|
|
}
|
|
|
|
if err := writeOutpoint(&b, &edgeInfo.ChannelPoint); err != nil {
|
|
return err
|
|
}
|
|
if err := binary.Write(&b, byteOrder, uint64(edgeInfo.Capacity)); err != nil {
|
|
return err
|
|
}
|
|
if _, err := b.Write(chanID[:]); err != nil {
|
|
return err
|
|
}
|
|
if _, err := b.Write(edgeInfo.ChainHash[:]); err != nil {
|
|
return err
|
|
}
|
|
|
|
return edgeIndex.Put(chanID[:], b.Bytes())
|
|
}
|
|
|
|
func fetchChanEdgeInfo(edgeIndex *bolt.Bucket,
|
|
chanID []byte) (ChannelEdgeInfo, error) {
|
|
|
|
edgeInfoBytes := edgeIndex.Get(chanID)
|
|
if edgeInfoBytes == nil {
|
|
return ChannelEdgeInfo{}, ErrEdgeNotFound
|
|
}
|
|
|
|
edgeInfoReader := bytes.NewReader(edgeInfoBytes)
|
|
return deserializeChanEdgeInfo(edgeInfoReader)
|
|
}
|
|
|
|
func deserializeChanEdgeInfo(r io.Reader) (ChannelEdgeInfo, error) {
|
|
var (
|
|
err error
|
|
edgeInfo ChannelEdgeInfo
|
|
)
|
|
|
|
if _, err := io.ReadFull(r, edgeInfo.NodeKey1Bytes[:]); err != nil {
|
|
return ChannelEdgeInfo{}, err
|
|
}
|
|
if _, err := io.ReadFull(r, edgeInfo.NodeKey2Bytes[:]); err != nil {
|
|
return ChannelEdgeInfo{}, err
|
|
}
|
|
if _, err := io.ReadFull(r, edgeInfo.BitcoinKey1Bytes[:]); err != nil {
|
|
return ChannelEdgeInfo{}, err
|
|
}
|
|
if _, err := io.ReadFull(r, edgeInfo.BitcoinKey2Bytes[:]); err != nil {
|
|
return ChannelEdgeInfo{}, err
|
|
}
|
|
|
|
edgeInfo.Features, err = wire.ReadVarBytes(r, 0, 900, "features")
|
|
if err != nil {
|
|
return ChannelEdgeInfo{}, err
|
|
}
|
|
|
|
proof := &ChannelAuthProof{}
|
|
|
|
proof.NodeSig1Bytes, err = wire.ReadVarBytes(r, 0, 80, "sigs")
|
|
if err != nil {
|
|
return ChannelEdgeInfo{}, err
|
|
}
|
|
proof.NodeSig2Bytes, err = wire.ReadVarBytes(r, 0, 80, "sigs")
|
|
if err != nil {
|
|
return ChannelEdgeInfo{}, err
|
|
}
|
|
proof.BitcoinSig1Bytes, err = wire.ReadVarBytes(r, 0, 80, "sigs")
|
|
if err != nil {
|
|
return ChannelEdgeInfo{}, err
|
|
}
|
|
proof.BitcoinSig2Bytes, err = wire.ReadVarBytes(r, 0, 80, "sigs")
|
|
if err != nil {
|
|
return ChannelEdgeInfo{}, err
|
|
}
|
|
|
|
if !proof.IsEmpty() {
|
|
edgeInfo.AuthProof = proof
|
|
}
|
|
|
|
edgeInfo.ChannelPoint = wire.OutPoint{}
|
|
if err := readOutpoint(r, &edgeInfo.ChannelPoint); err != nil {
|
|
return ChannelEdgeInfo{}, err
|
|
}
|
|
if err := binary.Read(r, byteOrder, &edgeInfo.Capacity); err != nil {
|
|
return ChannelEdgeInfo{}, err
|
|
}
|
|
if err := binary.Read(r, byteOrder, &edgeInfo.ChannelID); err != nil {
|
|
return ChannelEdgeInfo{}, err
|
|
}
|
|
|
|
if _, err := io.ReadFull(r, edgeInfo.ChainHash[:]); err != nil {
|
|
return ChannelEdgeInfo{}, err
|
|
}
|
|
|
|
return edgeInfo, nil
|
|
}
|
|
|
|
func putChanEdgePolicy(edges *bolt.Bucket, edge *ChannelEdgePolicy, from, to []byte) error {
|
|
var edgeKey [33 + 8]byte
|
|
copy(edgeKey[:], from)
|
|
byteOrder.PutUint64(edgeKey[33:], edge.ChannelID)
|
|
|
|
var b bytes.Buffer
|
|
|
|
err := wire.WriteVarBytes(&b, 0, edge.SigBytes)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
if err := binary.Write(&b, byteOrder, edge.ChannelID); err != nil {
|
|
return err
|
|
}
|
|
|
|
var scratch [8]byte
|
|
updateUnix := uint64(edge.LastUpdate.Unix())
|
|
byteOrder.PutUint64(scratch[:], updateUnix)
|
|
if _, err := b.Write(scratch[:]); err != nil {
|
|
return err
|
|
}
|
|
|
|
if err := binary.Write(&b, byteOrder, edge.Flags); err != nil {
|
|
return err
|
|
}
|
|
if err := binary.Write(&b, byteOrder, edge.TimeLockDelta); err != nil {
|
|
return err
|
|
}
|
|
if err := binary.Write(&b, byteOrder, uint64(edge.MinHTLC)); err != nil {
|
|
return err
|
|
}
|
|
if err := binary.Write(&b, byteOrder, uint64(edge.FeeBaseMSat)); err != nil {
|
|
return err
|
|
}
|
|
if err := binary.Write(&b, byteOrder, uint64(edge.FeeProportionalMillionths)); err != nil {
|
|
return err
|
|
}
|
|
|
|
if _, err := b.Write(to); err != nil {
|
|
return err
|
|
}
|
|
|
|
// Before we write out the new edge, we'll create a new entry in the
|
|
// update index in order to keep it fresh.
|
|
var indexKey [8 + 8]byte
|
|
copy(indexKey[:], scratch[:])
|
|
byteOrder.PutUint64(indexKey[8:], edge.ChannelID)
|
|
|
|
updateIndex, err := edges.CreateBucketIfNotExists(edgeUpdateIndexBucket)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
// If there was already an entry for this edge, then we'll need to
|
|
// delete the old one to ensure we don't leave around any after-images.
|
|
// An unknown policy value does not have a update time recorded, so
|
|
// it also does not need to be removed.
|
|
if edgeBytes := edges.Get(edgeKey[:]); edgeBytes != nil &&
|
|
!bytes.Equal(edgeBytes[:], unknownPolicy) {
|
|
|
|
// In order to delete the old entry, we'll need to obtain the
|
|
// *prior* update time in order to delete it. To do this, we'll
|
|
// create an offset to slice in. Starting backwards, we'll
|
|
// create an offset than puts us right after the flags
|
|
// variable:
|
|
//
|
|
// * pubkeySize + fee+policySize + timelockSize + flagSize
|
|
updateEnd := 33 + (8 * 3) + 2 + 1
|
|
updateStart := updateEnd - 8
|
|
oldUpdateTime := edgeBytes[updateStart:updateEnd]
|
|
|
|
var oldIndexKey [8 + 8]byte
|
|
copy(oldIndexKey[:], oldUpdateTime)
|
|
byteOrder.PutUint64(oldIndexKey[8:], edge.ChannelID)
|
|
|
|
if err := updateIndex.Delete(oldIndexKey[:]); err != nil {
|
|
return err
|
|
}
|
|
}
|
|
|
|
if err := updateIndex.Put(indexKey[:], nil); err != nil {
|
|
return err
|
|
}
|
|
|
|
return edges.Put(edgeKey[:], b.Bytes()[:])
|
|
}
|
|
|
|
// putChanEdgePolicyUnknown marks the edge policy as unknown
|
|
// in the edges bucket.
|
|
func putChanEdgePolicyUnknown(edges *bolt.Bucket, channelID uint64,
|
|
from []byte) error {
|
|
|
|
var edgeKey [33 + 8]byte
|
|
copy(edgeKey[:], from)
|
|
byteOrder.PutUint64(edgeKey[33:], channelID)
|
|
|
|
if edges.Get(edgeKey[:]) != nil {
|
|
return fmt.Errorf("Cannot write unknown policy for channel %v "+
|
|
" when there is already a policy present", channelID)
|
|
}
|
|
|
|
return edges.Put(edgeKey[:], unknownPolicy)
|
|
}
|
|
|
|
func fetchChanEdgePolicy(edges *bolt.Bucket, chanID []byte,
|
|
nodePub []byte, nodes *bolt.Bucket) (*ChannelEdgePolicy, error) {
|
|
|
|
var edgeKey [33 + 8]byte
|
|
copy(edgeKey[:], nodePub)
|
|
copy(edgeKey[33:], chanID[:])
|
|
|
|
edgeBytes := edges.Get(edgeKey[:])
|
|
if edgeBytes == nil {
|
|
return nil, ErrEdgeNotFound
|
|
}
|
|
|
|
// No need to deserialize unknown policy.
|
|
if bytes.Equal(edgeBytes[:], unknownPolicy) {
|
|
return nil, nil
|
|
}
|
|
|
|
edgeReader := bytes.NewReader(edgeBytes)
|
|
|
|
return deserializeChanEdgePolicy(edgeReader, nodes)
|
|
}
|
|
|
|
func fetchChanEdgePolicies(edgeIndex *bolt.Bucket, edges *bolt.Bucket,
|
|
nodes *bolt.Bucket, chanID []byte,
|
|
db *DB) (*ChannelEdgePolicy, *ChannelEdgePolicy, error) {
|
|
|
|
edgeInfo := edgeIndex.Get(chanID)
|
|
if edgeInfo == nil {
|
|
return nil, nil, ErrEdgeNotFound
|
|
}
|
|
|
|
// The first node is contained within the first half of the edge
|
|
// information. We only propagate the error here and below if it's
|
|
// something other than edge non-existence.
|
|
node1Pub := edgeInfo[:33]
|
|
edge1, err := fetchChanEdgePolicy(edges, chanID, node1Pub, nodes)
|
|
if err != nil {
|
|
return nil, nil, err
|
|
}
|
|
|
|
// As we may have a single direction of the edge but not the other,
|
|
// only fill in the database pointers if the edge is found.
|
|
if edge1 != nil {
|
|
edge1.db = db
|
|
edge1.Node.db = db
|
|
}
|
|
|
|
// Similarly, the second node is contained within the latter
|
|
// half of the edge information.
|
|
node2Pub := edgeInfo[33:67]
|
|
edge2, err := fetchChanEdgePolicy(edges, chanID, node2Pub, nodes)
|
|
if err != nil {
|
|
return nil, nil, err
|
|
}
|
|
|
|
if edge2 != nil {
|
|
edge2.db = db
|
|
edge2.Node.db = db
|
|
}
|
|
|
|
return edge1, edge2, nil
|
|
}
|
|
|
|
func deserializeChanEdgePolicy(r io.Reader,
|
|
nodes *bolt.Bucket) (*ChannelEdgePolicy, error) {
|
|
|
|
edge := &ChannelEdgePolicy{}
|
|
|
|
var err error
|
|
edge.SigBytes, err = wire.ReadVarBytes(r, 0, 80, "sig")
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
if err := binary.Read(r, byteOrder, &edge.ChannelID); err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
var scratch [8]byte
|
|
if _, err := r.Read(scratch[:]); err != nil {
|
|
return nil, err
|
|
}
|
|
unix := int64(byteOrder.Uint64(scratch[:]))
|
|
edge.LastUpdate = time.Unix(unix, 0)
|
|
|
|
if err := binary.Read(r, byteOrder, &edge.Flags); err != nil {
|
|
return nil, err
|
|
}
|
|
if err := binary.Read(r, byteOrder, &edge.TimeLockDelta); err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
var n uint64
|
|
if err := binary.Read(r, byteOrder, &n); err != nil {
|
|
return nil, err
|
|
}
|
|
edge.MinHTLC = lnwire.MilliSatoshi(n)
|
|
|
|
if err := binary.Read(r, byteOrder, &n); err != nil {
|
|
return nil, err
|
|
}
|
|
edge.FeeBaseMSat = lnwire.MilliSatoshi(n)
|
|
|
|
if err := binary.Read(r, byteOrder, &n); err != nil {
|
|
return nil, err
|
|
}
|
|
edge.FeeProportionalMillionths = lnwire.MilliSatoshi(n)
|
|
|
|
var pub [33]byte
|
|
if _, err := r.Read(pub[:]); err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
node, err := fetchLightningNode(nodes, pub[:])
|
|
if err != nil {
|
|
return nil, fmt.Errorf("unable to fetch node: %x, %v",
|
|
pub[:], err)
|
|
}
|
|
|
|
edge.Node = &node
|
|
return edge, nil
|
|
}
|