Network Resource Manager interface (#229)

* add resource manager interfaces

* add scope accessors to streams and conns

* add ResourceManager accessor to the Network interface

* allow initially unattached streams.

* introduce service scopes, canonicalize ownership interface through setters

* make system scope explicit

* make memory stat an int64

* make the system scope a generic resource scope, introduce the DMZ

* fix typo

Co-authored-by: Marten Seemann <martenseemann@gmail.com>

* fix typo

Co-authored-by: Marten Seemann <martenseemann@gmail.com>

* rename DMZ to transient scope, remove OpenConnection from PeerScope

* remove ncopy param from GrowBuffer

* remove protocols from OpenStream

The stream is unnegotiated state until the actual protocol has been determined.

* document nil receiver contract, fix protocol scope protocol accessor method

* remove nil receiver contract requirement

* flesh out stat struct

* turn resource manager scope accessors into viewers

* interface refiniments

1. Introduce transactions in all scopes
2. Limit the view of stream/connection scope for users, to avoid the Done footgun
3. Move OpenStream to the resource manager

* add Buffer interface

* fix typo

Co-authored-by: Marten Seemann <martenseemann@gmail.com>

* fix typo

Co-authored-by: Marten Seemann <martenseemann@gmail.com>

* fix typo

Co-authored-by: Marten Seemann <martenseemann@gmail.com>

* rename user scopes to plain names, management scopes as such

* rename BeginTxn to BeginTransaction

* RIP Buffers

* make ErrResourceLimitExceeded a temporary error; move rcmgr errors with the other errors.

* unexport TemporaryError

* null resource manager stub

* unexport the null stubs, make entry point a variable

* don't rely on typed nils but instead use actual null object instances

So that we don't confuse the hell out of everyone!

* add Scope to the CapableConn interface

* rename ConnectionScope to ConnScope for consistency

* fix typo

* rename ConnectionManagementScope to ConnManagementScope

* add the ConnManagementScope to Upgrader.Upgrade

* fix argument name

* godocs for ResourceManager

* introduce MemoryStatus indicator in ReserveMemory

* use uint8 for MemoryStatus

Co-authored-by: Marten Seemann <martenseemann@gmail.com>

* rework reservation interface to pass priority instead of returning memory status

so that we don't have to undo reservations if there is too much pressure.

* improve comment

* fix typo

* export the NullScope

* Stream.SetProtocol can return an error

It needs to attach the protocol to the protocol scope, which may fail.

* merge the mux package into network

* pass the PeerScope to Multiplexer.NetConn

* Update network/rcmgr.go

Co-authored-by: raulk <raul@protocol.ai>

* Update network/rcmgr.go

Co-authored-by: raulk <raul@protocol.ai>

* Update network/rcmgr.go

Co-authored-by: raulk <raul@protocol.ai>

* Update network/rcmgr.go

Co-authored-by: Adin Schmahmann <adin.schmahmann@gmail.com>

* remove reference to deprecated mux.MuxedConn

* rename transaction to span

* indicate bytes in ReserveMemory

* break ResourceManager View methods into Viewer interface.

* add experimental interface warning

Co-authored-by: Marten Seemann <martenseemann@gmail.com>
Co-authored-by: raulk <raul@protocol.ai>
Co-authored-by: Adin Schmahmann <adin.schmahmann@gmail.com>
This commit is contained in:
vyzo 2022-01-17 09:15:05 +02:00 committed by GitHub
parent ecd8d78e97
commit feb4853616
8 changed files with 469 additions and 98 deletions

View File

@ -5,99 +5,17 @@
package mux
import (
"context"
"errors"
"io"
"net"
"time"
"github.com/libp2p/go-libp2p-core/network"
)
// ErrReset is returned when reading or writing on a reset stream.
var ErrReset = errors.New("stream reset")
// Deprecated: use network.ErrReset instead.
var ErrReset = network.ErrReset
// Stream is a bidirectional io pipe within a connection.
type MuxedStream interface {
io.Reader
io.Writer
// Deprecated: use network.MuxedStream instead.
type MuxedStream = network.MuxedStream
// Close closes the stream.
//
// * Any buffered data for writing will be flushed.
// * Future reads will fail.
// * Any in-progress reads/writes will be interrupted.
//
// Close may be asynchronous and _does not_ guarantee receipt of the
// data.
//
// Close closes the stream for both reading and writing.
// Close is equivalent to calling `CloseRead` and `CloseWrite`. Importantly, Close will not wait for any form of acknowledgment.
// If acknowledgment is required, the caller must call `CloseWrite`, then wait on the stream for a response (or an EOF),
// then call Close() to free the stream object.
//
// When done with a stream, the user must call either Close() or `Reset()` to discard the stream, even after calling `CloseRead` and/or `CloseWrite`.
io.Closer
// Deprecated: use network.MuxedConn instead.
type MuxedConn = network.MuxedConn
// CloseWrite closes the stream for writing but leaves it open for
// reading.
//
// CloseWrite does not free the stream, users must still call Close or
// Reset.
CloseWrite() error
// CloseRead closes the stream for reading but leaves it open for
// writing.
//
// When CloseRead is called, all in-progress Read calls are interrupted with a non-EOF error and
// no further calls to Read will succeed.
//
// The handling of new incoming data on the stream after calling this function is implementation defined.
//
// CloseRead does not free the stream, users must still call Close or
// Reset.
CloseRead() error
// Reset closes both ends of the stream. Use this to tell the remote
// side to hang up and go away.
Reset() error
SetDeadline(time.Time) error
SetReadDeadline(time.Time) error
SetWriteDeadline(time.Time) error
}
// NoopHandler do nothing. Resets streams as soon as they are opened.
var NoopHandler = func(s MuxedStream) { s.Reset() }
// MuxedConn represents a connection to a remote peer that has been
// extended to support stream multiplexing.
//
// A MuxedConn allows a single net.Conn connection to carry many logically
// independent bidirectional streams of binary data.
//
// Together with network.ConnSecurity, MuxedConn is a component of the
// transport.CapableConn interface, which represents a "raw" network
// connection that has been "upgraded" to support the libp2p capabilities
// of secure communication and stream multiplexing.
type MuxedConn interface {
// Close closes the stream muxer and the the underlying net.Conn.
io.Closer
// IsClosed returns whether a connection is fully closed, so it can
// be garbage collected.
IsClosed() bool
// OpenStream creates a new stream.
OpenStream(context.Context) (MuxedStream, error)
// AcceptStream accepts a stream opened by the other side.
AcceptStream() (MuxedStream, error)
}
// Multiplexer wraps a net.Conn with a stream multiplexing
// implementation and returns a MuxedConn that supports opening
// multiple streams over the underlying net.Conn
type Multiplexer interface {
// NewConn constructs a new connection
NewConn(c net.Conn, isServer bool) (MuxedConn, error)
}
// Deprecated: use network.Multiplexer instead.
type Multiplexer = network.Multiplexer

View File

@ -20,6 +20,7 @@ type Conn interface {
ConnSecurity
ConnMultiaddrs
ConnStat
ConnScoper
// ID returns an identifier that uniquely identifies this Conn within this
// host, during this run. Connection IDs may repeat across restarts.
@ -65,3 +66,10 @@ type ConnStat interface {
// Stat stores metadata pertaining to this conn.
Stat() ConnStats
}
// ConnScoper is the interface that one can mix into a connection interface to give it a resource
// management scope
type ConnScoper interface {
// Scope returns the user view of this connection's resource scope
Scope() ConnScope
}

View File

@ -1,6 +1,17 @@
package network
import "errors"
import (
"errors"
"net"
)
type temporaryError string
func (e temporaryError) Error() string { return string(e) }
func (e temporaryError) Temporary() bool { return true }
func (e temporaryError) Timeout() bool { return false }
var _ net.Error = temporaryError("")
// ErrNoRemoteAddrs is returned when there are no addresses associated with a peer during a dial.
var ErrNoRemoteAddrs = errors.New("no remote addresses")
@ -12,3 +23,11 @@ var ErrNoConn = errors.New("no usable connection to peer")
// ErrTransientConn is returned when attempting to open a stream to a peer with only a transient
// connection, without specifying the UseTransient option.
var ErrTransientConn = errors.New("transient connection to peer")
// ErrResourceLimitExceeded is returned when attempting to perform an operation that would
// exceed system resource limits.
var ErrResourceLimitExceeded = temporaryError("resource limit exceeded")
// ErrResourceScopeClosed is returned when attemptig to reserve resources in a closed resource
// scope.
var ErrResourceScopeClosed = errors.New("resource scope closed")

95
core/network/mux.go Normal file
View File

@ -0,0 +1,95 @@
package network
import (
"context"
"errors"
"io"
"net"
"time"
)
// ErrReset is returned when reading or writing on a reset stream.
var ErrReset = errors.New("stream reset")
// MuxedStream is a bidirectional io pipe within a connection.
type MuxedStream interface {
io.Reader
io.Writer
// Close closes the stream.
//
// * Any buffered data for writing will be flushed.
// * Future reads will fail.
// * Any in-progress reads/writes will be interrupted.
//
// Close may be asynchronous and _does not_ guarantee receipt of the
// data.
//
// Close closes the stream for both reading and writing.
// Close is equivalent to calling `CloseRead` and `CloseWrite`. Importantly, Close will not wait for any form of acknowledgment.
// If acknowledgment is required, the caller must call `CloseWrite`, then wait on the stream for a response (or an EOF),
// then call Close() to free the stream object.
//
// When done with a stream, the user must call either Close() or `Reset()` to discard the stream, even after calling `CloseRead` and/or `CloseWrite`.
io.Closer
// CloseWrite closes the stream for writing but leaves it open for
// reading.
//
// CloseWrite does not free the stream, users must still call Close or
// Reset.
CloseWrite() error
// CloseRead closes the stream for reading but leaves it open for
// writing.
//
// When CloseRead is called, all in-progress Read calls are interrupted with a non-EOF error and
// no further calls to Read will succeed.
//
// The handling of new incoming data on the stream after calling this function is implementation defined.
//
// CloseRead does not free the stream, users must still call Close or
// Reset.
CloseRead() error
// Reset closes both ends of the stream. Use this to tell the remote
// side to hang up and go away.
Reset() error
SetDeadline(time.Time) error
SetReadDeadline(time.Time) error
SetWriteDeadline(time.Time) error
}
// MuxedConn represents a connection to a remote peer that has been
// extended to support stream multiplexing.
//
// A MuxedConn allows a single net.Conn connection to carry many logically
// independent bidirectional streams of binary data.
//
// Together with network.ConnSecurity, MuxedConn is a component of the
// transport.CapableConn interface, which represents a "raw" network
// connection that has been "upgraded" to support the libp2p capabilities
// of secure communication and stream multiplexing.
type MuxedConn interface {
// Close closes the stream muxer and the the underlying net.Conn.
io.Closer
// IsClosed returns whether a connection is fully closed, so it can
// be garbage collected.
IsClosed() bool
// OpenStream creates a new stream.
OpenStream(context.Context) (MuxedStream, error)
// AcceptStream accepts a stream opened by the other side.
AcceptStream() (MuxedStream, error)
}
// Multiplexer wraps a net.Conn with a stream multiplexing
// implementation and returns a MuxedConn that supports opening
// multiple streams over the underlying net.Conn
type Multiplexer interface {
// NewConn constructs a new connection
NewConn(c net.Conn, isServer bool, scope PeerScope) (MuxedConn, error)
}

View File

@ -145,6 +145,9 @@ type Network interface {
// listens. It expands "any interface" addresses (/ip4/0.0.0.0, /ip6/::) to
// use the known local interfaces.
InterfaceListenAddresses() ([]ma.Multiaddr, error)
// ResourceManager returns the ResourceManager associated with this network
ResourceManager() ResourceManager
}
// Dialer represents a service that can dial out to peers

326
core/network/rcmgr.go Normal file
View File

@ -0,0 +1,326 @@
package network
import (
"github.com/libp2p/go-libp2p-core/peer"
"github.com/libp2p/go-libp2p-core/protocol"
)
// ResourceManager is the interface to the network resource management subsystem.
// The ResourceManager tracks and accounts for resource usage in the stack, from the internals
// to the application, and provides a mechanism to limit resource usage according to a user
// configurable policy.
//
// Resource Management through the ResourceManager is based on the concept of Resource
// Management Scopes, whereby resource usage is constrained by a DAG of scopes,
// The following diagram illustrates the structure of the resource constraint DAG:
// System
// +------------> Transient.............+................+
// | . .
// +------------> Service------------- . ----------+ .
// | . | .
// +-------------> Protocol----------- . ----------+ .
// | . | .
// +--------------> Peer \ | .
// +------------> Connection | .
// | \ \
// +---------------------------> Stream
//
// The basic resources accounted by the ResourceManager include memory, streams, connections,
// and file descriptors. These account for both space and time used by
// the stack, as each resource has a direct effect on the system
// availability and performance.
//
// The modus operandi of the resource manager is to restrict resource usage at the time of
// reservation. When a component of the stack needs to use a resource, it reserves it in the
// appropriate scope. The resource manager gates the reservation against the scope applicable
// limits; if the limit is exceeded, then an error (wrapping ErrResourceLimitExceeded) and it
// is up the component to act accordingly. At the lower levels of the stack, this will normally
// signal a failure of some sorts, like failing to opening a stream or a connection, which will
// propagate to the programmer. Some components may be able to handle resource reservation failure
// more gracefully; for instance a muxer trying to grow a buffer for a window change, will simply
// retain the existing window size and continue to operate normally albeit with some degraded
// throughput.
// All resources reserved in some scope are released when the scope is closed. For low level
// scopes, mainly Connection and Stream scopes, this happens when the connection or stream is
// closed.
//
// Service programmers will typically use the resource manager to reserve memory
// for their subsystem.
// This happens with two avenues: the programmer can attach a stream to a service, whereby
// resources reserved by the stream are automatically accounted in the service budget; or the
// programmer may directly interact with the service scope, by using ViewService through the
// resource manager interface.
//
// Application programmers can also directly reserve memory in some applicable scope. In order
// to facilitate control flow delimited resource accounting, all scopes defined in the system
// allow for the user to create spans. Spans are temporary scopes rooted at some
// other scope and release their resources when the programmer is done with them. Span
// scopes can form trees, with nested spans.
//
// Typical Usage:
// - Low level components of the system (transports, muxers) all have access to the resource
// manager and create connection and stream scopes through it. These scopes are accessible
// to the user, albeit with a narrower interface, through Conn and Stream objects who have
// a Scope method.
// - Services typically center around streams, where the programmer can attach streams to a
// particular service. They can also directly reserve memory for a service by accessing the
// service scope using the ResourceManager interface.
// - Applications that want to account for their network resource usage can reserve memory,
// typically using a span, directly in the System or a Service scope; they can also
// opt to use appropriate steam scopes for streams that they create or own.
//
// User Serviceable Parts: the user has the option to specify their own implementation of the
// interface. We provide a canonical implementation in the go-libp2p-resource-manager package.
// The user of that package can specify limits for the various scopes, which can be static
// or dynamic.
//
// WARNING The ResourceManager interface is considered experimental and subject to change
// in subsequent releases.
type ResourceManager interface {
ResourceScopeViewer
// OpenConnection creates a new connection scope not yet associated with any peer; the connection
// is scoped at the transient scope.
// The caller owns the returned scope and is responsible for calling Done in order to signify
// the end of the scope's span.
OpenConnection(dir Direction, usefd bool) (ConnManagementScope, error)
// OpenStream creates a new stream scope, initially unnegotiated.
// An unnegotiated stream will be initially unattached to any protocol scope
// and constrained by the transient scope.
// The caller owns the returned scope and is responsible for calling Done in order to signify
// the end of th scope's span.
OpenStream(p peer.ID, dir Direction) (StreamManagementScope, error)
// Close closes the resource manager
Close() error
}
// ResourceScopeViewer is a mixin interface providing view methods for accessing top level
// scopes.
type ResourceScopeViewer interface {
// ViewSystem views the system wide resource scope.
// The system scope is the top level scope that accounts for global
// resource usage at all levels of the system. This scope constrains all
// other scopes and institutes global hard limits.
ViewSystem(func(ResourceScope) error) error
// ViewTransient views the transient (DMZ) resource scope.
// The transient scope accounts for resources that are in the process of
// full establishment. For instance, a new connection prior to the
// handshake does not belong to any peer, but it still needs to be
// constrained as this opens an avenue for attacks in transient resource
// usage. Similarly, a stream that has not negotiated a protocol yet is
// constrained by the transient scope.
ViewTransient(func(ResourceScope) error) error
// ViewService retrieves a service-specific scope.
ViewService(string, func(ServiceScope) error) error
// ViewProtocol views the resource management scope for a specific protocol.
ViewProtocol(protocol.ID, func(ProtocolScope) error) error
// ViewPeer views the resource management scope for a specific peer.
ViewPeer(peer.ID, func(PeerScope) error) error
}
const (
// ReservationPriorityLow is a reservation priority that indicates a reservation if the scope
// memory utilization is at 40% or less.
ReservationPriorityLow uint8 = 101
// Reservation PriorityMedium is a reservation priority that indicates a reservation if the scope
// memory utilization is at 60% or less.
ReservationPriorityMedium uint8 = 152
// ReservationPriorityHigh is a reservation prioirity that indicates a reservation if the scope
// memory utilization is at 80% or less.
ReservationPriorityHigh uint8 = 203
// ReservationPriorityAlways is a reservation priority that indicates a reservation if there is
// enough memory, regardless of scope utilization.
ReservationPriorityAlways uint8 = 255
)
// ResourceScope is the interface for all scopes.
type ResourceScope interface {
// ReserveMemory reserves memory/buffer space in the scope; the unit is bytes.
//
// If ReserveMemory returns an error, then no memory was reserved and the caller should handle
// the failure condition.
//
// The priority argument indicates the priority of the memory reservation. A reservation
// will fail if the available memory is less than (1+prio)/256 of the scope limit, providing
// a mechanism to gracefully handle optional reservations that might overload the system.
// For instance, a muxer growing a window buffer will use a low priority and only grow the buffer
// if there is no memory pressure in the system.
//
// The are 4 predefined priority levels, Low, Medium, High and Always,
// capturing common patterns, but the user is free to use any granularity applicable to his case.
ReserveMemory(size int, prio uint8) error
// ReleaseMemory explicitly releases memory previously reserved with ReserveMemory
ReleaseMemory(size int)
// Stat retrieves current resource usage for the scope.
Stat() ScopeStat
// BeginSpan creates a new span scope rooted at this scope
BeginSpan() (ResourceScopeSpan, error)
}
// ResourceScopeSpan is a ResourceScope with a delimited span.
// Span scopes are control flow delimited and release all their associated resources
// when the programmer calls Done.
//
// Example:
// s, err := someScope.BeginSpan()
// if err != nil { ... }
// defer s.Done()
//
// if err := s.ReserveMemory(...); err != nil { ... }
// // ... use memory
type ResourceScopeSpan interface {
ResourceScope
// Done ends the span and releases associated resources.
Done()
}
// ServiceScope is the interface for service resource scopes
type ServiceScope interface {
ResourceScope
// Name returns the name of this service
Name() string
}
// ProtocolScope is the interface for protocol resource scopes.
type ProtocolScope interface {
ResourceScope
// Protocol returns the protocol for this scope
Protocol() protocol.ID
}
// PeerScope is the interface for peer resource scopes.
type PeerScope interface {
ResourceScope
// Peer returns the peer ID for this scope
Peer() peer.ID
}
// ConnManagementScope is the low level interface for connection resource scopes.
// This interface is used by the low level components of the system who create and own
// the span of a connection scope.
type ConnManagementScope interface {
ResourceScopeSpan
// PeerScope returns the peer scope associated with this connection.
// It returns nil if the connection is not yet asociated with any peer.
PeerScope() PeerScope
// SetPeer sets the peer for a previously unassociated connection
SetPeer(peer.ID) error
}
// ConnScope is the user view of a connection scope
type ConnScope interface {
ResourceScope
}
// StreamManagementScope is the interface for stream resource scopes.
// This interface is used by the low level components of the system who create and own
// the span of a stream scope.
type StreamManagementScope interface {
ResourceScopeSpan
// ProtocolScope returns the protocol resource scope associated with this stream.
// It returns nil if the stream is not associated with any protocol scope.
ProtocolScope() ProtocolScope
// SetProtocol sets the protocol for a previously unnegotiated stream
SetProtocol(proto protocol.ID) error
// ServiceScope returns the service owning the stream, if any.
ServiceScope() ServiceScope
// SetService sets the service owning this stream.
SetService(srv string) error
// PeerScope returns the peer resource scope associated with this stream.
PeerScope() PeerScope
}
// StreamScope is the user view of a StreamScope.
type StreamScope interface {
ResourceScope
// SetService sets the service owning this stream.
SetService(srv string) error
}
// ScopeStat is a struct containing resource accounting information.
type ScopeStat struct {
NumStreamsInbound int
NumStreamsOutbound int
NumConnsInbound int
NumConnsOutbound int
NumFD int
Memory int64
}
// NullResourceManager is a stub for tests and initialization of default values
var NullResourceManager ResourceManager = &nullResourceManager{}
type nullResourceManager struct{}
type nullScope struct{}
var _ ResourceScope = (*nullScope)(nil)
var _ ResourceScopeSpan = (*nullScope)(nil)
var _ ServiceScope = (*nullScope)(nil)
var _ ProtocolScope = (*nullScope)(nil)
var _ PeerScope = (*nullScope)(nil)
var _ ConnManagementScope = (*nullScope)(nil)
var _ ConnScope = (*nullScope)(nil)
var _ StreamManagementScope = (*nullScope)(nil)
var _ StreamScope = (*nullScope)(nil)
// NullScope is a stub for tests and initialization of default values
var NullScope = &nullScope{}
func (n *nullResourceManager) ViewSystem(f func(ResourceScope) error) error {
return f(NullScope)
}
func (n *nullResourceManager) ViewTransient(f func(ResourceScope) error) error {
return f(NullScope)
}
func (n *nullResourceManager) ViewService(svc string, f func(ServiceScope) error) error {
return f(NullScope)
}
func (n *nullResourceManager) ViewProtocol(p protocol.ID, f func(ProtocolScope) error) error {
return f(NullScope)
}
func (n *nullResourceManager) ViewPeer(p peer.ID, f func(PeerScope) error) error {
return f(NullScope)
}
func (n *nullResourceManager) OpenConnection(dir Direction, usefd bool) (ConnManagementScope, error) {
return NullScope, nil
}
func (n *nullResourceManager) OpenStream(p peer.ID, dir Direction) (StreamManagementScope, error) {
return NullScope, nil
}
func (n *nullResourceManager) Close() error {
return nil
}
func (n *nullScope) ReserveMemory(size int, prio uint8) error { return nil }
func (n *nullScope) ReleaseMemory(size int) {}
func (n *nullScope) Stat() ScopeStat { return ScopeStat{} }
func (n *nullScope) BeginSpan() (ResourceScopeSpan, error) { return NullScope, nil }
func (n *nullScope) Done() {}
func (n *nullScope) Name() string { return "" }
func (n *nullScope) Protocol() protocol.ID { return "" }
func (n *nullScope) Peer() peer.ID { return "" }
func (n *nullScope) PeerScope() PeerScope { return NullScope }
func (n *nullScope) SetPeer(peer.ID) error { return nil }
func (n *nullScope) ProtocolScope() ProtocolScope { return NullScope }
func (n *nullScope) SetProtocol(proto protocol.ID) error { return nil }
func (n *nullScope) ServiceScope() ServiceScope { return NullScope }
func (n *nullScope) SetService(srv string) error { return nil }

View File

@ -1,7 +1,6 @@
package network
import (
"github.com/libp2p/go-libp2p-core/mux"
"github.com/libp2p/go-libp2p-core/protocol"
)
@ -11,18 +10,21 @@ import (
//
// Streams are backed by a multiplexer underneath the hood.
type Stream interface {
mux.MuxedStream
MuxedStream
// ID returns an identifier that uniquely identifies this Stream within this
// host, during this run. Stream IDs may repeat across restarts.
ID() string
Protocol() protocol.ID
SetProtocol(id protocol.ID)
SetProtocol(id protocol.ID) error
// Stat returns metadata pertaining to this stream.
Stat() Stats
// Conn returns the connection this stream is part of.
Conn() Conn
// Scope returns the user's view of this stream's resource scope
Scope() StreamScope
}

View File

@ -6,7 +6,6 @@ import (
"context"
"net"
"github.com/libp2p/go-libp2p-core/mux"
"github.com/libp2p/go-libp2p-core/network"
"github.com/libp2p/go-libp2p-core/peer"
@ -26,9 +25,10 @@ import (
// CapableConn provides accessors for the local and remote multiaddrs used to
// establish the connection and an accessor for the underlying Transport.
type CapableConn interface {
mux.MuxedConn
network.MuxedConn
network.ConnSecurity
network.ConnMultiaddrs
network.ConnScoper
// Transport returns the transport to which this connection belongs.
Transport() Transport
@ -112,5 +112,5 @@ type Upgrader interface {
// UpgradeListener upgrades the passed multiaddr-net listener into a full libp2p-transport listener.
UpgradeListener(Transport, manet.Listener) Listener
// Upgrade upgrades the multiaddr/net connection into a full libp2p-transport connection.
Upgrade(ctx context.Context, t Transport, maconn manet.Conn, dir network.Direction, p peer.ID) (CapableConn, error)
Upgrade(ctx context.Context, t Transport, maconn manet.Conn, dir network.Direction, p peer.ID, scope network.ConnManagementScope) (CapableConn, error)
}