netlink

package
v0.3.3 Latest Latest
Warning

This package is not in the latest version of its module.

Go to latest
Published: Jul 3, 2015 License: Apache-2.0, Apache-2.0 Imports: 8 Imported by: 0

README

Build Status GoDoc

The netlink package provides a simple netlink library for go. Netlink is the interface a user-space program in linux uses to communicate with the kernel. It can be used to add and remove interfaces, set ip addresses and routes, and configure ipsec. Netlink communication requires elevated privileges, so in most cases this code needs to be run as root. Since low-level netlink messages are inscrutable at best, the library attempts to provide an api that is loosely modeled on the CLI provied by iproute2. Actions like ip link add will be accomplished via a similarly named function like AddLink(). This library began its life as a fork of the netlink functionality in docker/libcontainer but was heavily rewritten to improve testability, performance, and to add new functionality like ipsec xfrm handling.

Local Build and Test

You can use go get command:

go get github.com/vishvananda/netlink

Testing dependencies:

go get github.com/vishvananda/netns

Testing (requires root):

sudo -E go test github.com/vishvananda/netlink

Examples

Add a new bridge and add eth1 into it:

package main

import (
    "net"
    "github.com/vishvananda/netlink"
)

func main() {
    mybridge := &netlink.Bridge{netlink.LinkAttrs{Name: "foo"}}
    _ := netlink.LinkAdd(mybridge)
    eth1, _ := netlink.LinkByName("eth1")
    netlink.LinkSetMaster(eth1, mybridge)
}

Add a new ip address to loopback:

package main

import (
    "net"
    "github.com/vishvananda/netlink"
)

func main() {
    lo, _ := netlink.LinkByName("lo")
    addr, _ := netlink.ParseAddr("169.254.169.254/32")
    netlink.AddrAdd(lo, addr)
}

Future Work

Many pieces of netlink are not yet fully supported in the high-level interface. Aspects of virtually all of the high-level objects don't exist. Many of the underlying primitives are there, so its a matter of putting the right fields into the high-level objects and making sure that they are serialized and deserialized correctly in the Add and List methods.

There are also a few pieces of low level netlink functionality that still need to be implemented. Routing rules are not in place and some of the more advanced link types. Hopefully there is decent structure and testing in place to make these fairly straightforward to add.

Documentation

Overview

Package netlink provides a simple library for netlink. Netlink is the interface a user-space program in linux uses to communicate with the kernel. It can be used to add and remove interfaces, set up ip addresses and routes, and confiugre ipsec. Netlink communication requires elevated privileges, so in most cases this code needs to be run as root. The low level primitives for netlink are contained in the nl subpackage. This package attempts to provide a high-level interface that is loosly modeled on the iproute2 cli.

Index

Constants

View Source
const (
	NDA_UNSPEC = iota
	NDA_DST
	NDA_LLADDR
	NDA_CACHEINFO
	NDA_PROBES
	NDA_VLAN
	NDA_PORT
	NDA_VNI
	NDA_IFINDEX
	NDA_MAX = NDA_IFINDEX
)
View Source
const (
	NUD_NONE       = 0x00
	NUD_INCOMPLETE = 0x01
	NUD_REACHABLE  = 0x02
	NUD_STALE      = 0x04
	NUD_DELAY      = 0x08
	NUD_PROBE      = 0x10
	NUD_FAILED     = 0x20
	NUD_NOARP      = 0x40
	NUD_PERMANENT  = 0x80
)

Neighbor Cache Entry States.

View Source
const (
	NTF_USE    = 0x01
	NTF_SELF   = 0x02
	NTF_MASTER = 0x04
	NTF_PROXY  = 0x08
	NTF_ROUTER = 0x80
)

Neighbor Flags

View Source
const (
	// Family type definitions
	FAMILY_ALL = nl.FAMILY_ALL
	FAMILY_V4  = nl.FAMILY_V4
	FAMILY_V6  = nl.FAMILY_V6
)

Variables

This section is empty.

Functions

func AddrAdd

func AddrAdd(link Link, addr *Addr) error

AddrAdd will add an IP address to a link device. Equivalent to: `ip addr add $addr dev $link`

func AddrDel

func AddrDel(link Link, addr *Addr) error

AddrDel will delete an IP address from a link device. Equivalent to: `ip addr del $addr dev $link`

func LinkAdd

func LinkAdd(link Link) error

LinkAdd adds a new link device. The type and features of the device are taken fromt the parameters in the link object. Equivalent to: `ip link add $link`

func LinkDel

func LinkDel(link Link) error

LinkDel deletes link device. Either Index or Name must be set in the link object for it to be deleted. The other values are ignored. Equivalent to: `ip link del $link`

func LinkSetDown

func LinkSetDown(link Link) error

LinkSetUp disables link device. Equivalent to: `ip link set $link down`

func LinkSetFastLeave

func LinkSetFastLeave(link Link, mode bool) error

func LinkSetFlood

func LinkSetFlood(link Link, mode bool) error

func LinkSetGuard

func LinkSetGuard(link Link, mode bool) error

func LinkSetHairpin

func LinkSetHairpin(link Link, mode bool) error

func LinkSetHardwareAddr

func LinkSetHardwareAddr(link Link, hwaddr net.HardwareAddr) error

LinkSetHardwareAddr sets the hardware address of the link device. Equivalent to: `ip link set $link address $hwaddr`

func LinkSetLearning

func LinkSetLearning(link Link, mode bool) error

func LinkSetMTU

func LinkSetMTU(link Link, mtu int) error

LinkSetMTU sets the mtu of the link device. Equivalent to: `ip link set $link mtu $mtu`

func LinkSetMaster

func LinkSetMaster(link Link, master *Bridge) error

LinkSetMaster sets the master of the link device. Equivalent to: `ip link set $link master $master`

func LinkSetMasterByIndex

func LinkSetMasterByIndex(link Link, masterIndex int) error

LinkSetMasterByIndex sets the master of the link device. Equivalent to: `ip link set $link master $master`

func LinkSetName

func LinkSetName(link Link, name string) error

LinkSetName sets the name of the link device. Equivalent to: `ip link set $link name $name`

func LinkSetNsFd

func LinkSetNsFd(link Link, fd int) error

LinkSetNsPid puts the device into a new network namespace. The fd must be an open file descriptor to a network namespace. Similar to: `ip link set $link netns $ns`

func LinkSetNsPid

func LinkSetNsPid(link Link, nspid int) error

LinkSetNsPid puts the device into a new network namespace. The pid must be a pid of a running process. Equivalent to: `ip link set $link netns $pid`

func LinkSetRootBlock

func LinkSetRootBlock(link Link, mode bool) error

func LinkSetUp

func LinkSetUp(link Link) error

LinkSetUp enables the link device. Equivalent to: `ip link set $link up`

func NeighAdd

func NeighAdd(neigh *Neigh) error

NeighAdd will add an IP to MAC mapping to the ARP table Equivalent to: `ip neigh add ....`

func NeighAppend

func NeighAppend(neigh *Neigh) error

NeighAppend will append an entry to FDB Equivalent to: `bridge fdb append...`

func NeighDel

func NeighDel(neigh *Neigh) error

NeighDel will delete an IP address from a link device. Equivalent to: `ip addr del $addr dev $link`

func NeighSet

func NeighSet(neigh *Neigh) error

NeighAdd will add or replace an IP to MAC mapping to the ARP table Equivalent to: `ip neigh replace....`

func NewIPNet

func NewIPNet(ip net.IP) *net.IPNet

NewIPNet generates an IPNet from an ip address using a netmask of 32.

func ParseIPNet

func ParseIPNet(s string) (*net.IPNet, error)

ParseIPNet parses a string in ip/net format and returns a net.IPNet. This is valuable because addresses in netlink are often IPNets and ParseCIDR returns an IPNet with the IP part set to the base IP of the range.

func RouteAdd

func RouteAdd(route *Route) error

RouteAdd will add a route to the system. Equivalent to: `ip route add $route`

func RouteDel

func RouteDel(route *Route) error

RouteAdd will delete a route from the system. Equivalent to: `ip route del $route`

func XfrmPolicyAdd

func XfrmPolicyAdd(policy *XfrmPolicy) error

XfrmPolicyAdd will add an xfrm policy to the system. Equivalent to: `ip xfrm policy add $policy`

func XfrmPolicyDel

func XfrmPolicyDel(policy *XfrmPolicy) error

XfrmPolicyDel will delete an xfrm policy from the system. Note that the Tmpls are ignored when matching the policy to delete. Equivalent to: `ip xfrm policy del $policy`

func XfrmStateAdd

func XfrmStateAdd(state *XfrmState) error

XfrmStateAdd will add an xfrm state to the system. Equivalent to: `ip xfrm state add $state`

func XfrmStateDel

func XfrmStateDel(state *XfrmState) error

XfrmStateDel will delete an xfrm state from the system. Note that the Algos are ignored when matching the state to delete. Equivalent to: `ip xfrm state del $state`

Types

type Addr

type Addr struct {
	*net.IPNet
	Label string
}

Addr represents an IP address from netlink. Netlink ip addresses include a mask, so it stores the address as a net.IPNet.

func AddrList

func AddrList(link Link, family int) ([]Addr, error)

AddrList gets a list of IP addresses in the system. Equivalent to: `ip addr show`. The list can be filtered by link and ip family.

func ParseAddr

func ParseAddr(s string) (*Addr, error)

ParseAddr parses the string representation of an address in the form $ip/$netmask $label. The label portion is optional

func (Addr) Equal

func (a Addr) Equal(x Addr) bool

Equal returns true if both Addrs have the same net.IPNet value.

func (Addr) String

func (addr Addr) String() string

String returns $ip/$netmask $label

type Bridge

type Bridge struct {
	LinkAttrs
}

Bridge links are simple linux bridges

func (*Bridge) Attrs

func (bridge *Bridge) Attrs() *LinkAttrs

func (*Bridge) Type

func (bridge *Bridge) Type() string

type Device

type Device struct {
	LinkAttrs
}

Device links cannot be created via netlink. These links are links created by udev like 'lo' and 'etho0'

func (*Device) Attrs

func (device *Device) Attrs() *LinkAttrs

func (*Device) Type

func (device *Device) Type() string

type Dir

type Dir uint8

Dir is an enum representing an ipsec template direction.

const (
	XFRM_DIR_IN Dir = iota
	XFRM_DIR_OUT
	XFRM_DIR_FWD
	XFRM_SOCKET_IN
	XFRM_SOCKET_OUT
	XFRM_SOCKET_FWD
)

func (Dir) String

func (d Dir) String() string

type Dummy

type Dummy struct {
	LinkAttrs
}

Dummy links are dummy ethernet devices

func (*Dummy) Attrs

func (dummy *Dummy) Attrs() *LinkAttrs

func (*Dummy) Type

func (dummy *Dummy) Type() string

type EncapType

type EncapType uint8

EncapType is an enum representing an ipsec template direction.

const (
	XFRM_ENCAP_ESPINUDP_NONIKE EncapType = iota + 1
	XFRM_ENCAP_ESPINUDP
)

func (EncapType) String

func (e EncapType) String() string

type Generic

type Generic struct {
	LinkAttrs
	LinkType string
}

Generic links represent types that are not currently understood by this netlink library.

func (*Generic) Attrs

func (generic *Generic) Attrs() *LinkAttrs

func (*Generic) Type

func (generic *Generic) Type() string

type IPVlan

type IPVlan struct {
	LinkAttrs
	Mode IPVlanMode
}

func (*IPVlan) Attrs

func (ipvlan *IPVlan) Attrs() *LinkAttrs

func (*IPVlan) Type

func (ipvlan *IPVlan) Type() string

type IPVlanMode

type IPVlanMode uint16
const (
	IPVLAN_MODE_L2 IPVlanMode = iota
	IPVLAN_MODE_L3
	IPVLAN_MODE_MAX
)
type Link interface {
	Attrs() *LinkAttrs
	Type() string
}

Link represents a link device from netlink. Shared link attributes like name may be retrieved using the Attrs() method. Unique data can be retrieved by casting the object to the proper type.

func LinkByIndex

func LinkByIndex(index int) (Link, error)

LinkByIndex finds a link by index and returns a pointer to the object.

func LinkByName

func LinkByName(name string) (Link, error)

LinkByName finds a link by name and returns a pointer to the object.

func LinkList() ([]Link, error)

LinkList gets a list of link devices. Equivalent to: `ip link show`

type LinkAttrs

type LinkAttrs struct {
	Index        int
	MTU          int
	TxQLen       uint32 // Transmit Queue Length
	Name         string
	HardwareAddr net.HardwareAddr
	Flags        net.Flags
	ParentIndex  int         // index of the parent link device
	MasterIndex  int         // must be the index of a bridge
	Namespace    interface{} // nil | NsPid | NsFd
}

LinkAttrs represents data shared by most link types

type Macvlan

type Macvlan struct {
	LinkAttrs
	Mode MacvlanMode
}

Macvlan links have ParentIndex set in their Attrs()

func (*Macvlan) Attrs

func (macvlan *Macvlan) Attrs() *LinkAttrs

func (*Macvlan) Type

func (macvlan *Macvlan) Type() string

type MacvlanMode

type MacvlanMode uint16
const (
	MACVLAN_MODE_DEFAULT MacvlanMode = iota
	MACVLAN_MODE_PRIVATE
	MACVLAN_MODE_VEPA
	MACVLAN_MODE_BRIDGE
	MACVLAN_MODE_PASSTHRU
	MACVLAN_MODE_SOURCE
)

type Mode

type Mode uint8

Mode is an enum representing an ipsec transport.

const (
	XFRM_MODE_TRANSPORT Mode = iota
	XFRM_MODE_TUNNEL
	XFRM_MODE_ROUTEOPTIMIZATION
	XFRM_MODE_IN_TRIGGER
	XFRM_MODE_BEET
	XFRM_MODE_MAX
)

func (Mode) String

func (m Mode) String() string

type Ndmsg

type Ndmsg struct {
	Family uint8
	Index  uint32
	State  uint16
	Flags  uint8
	Type   uint8
}

func (*Ndmsg) Len

func (msg *Ndmsg) Len() int

func (*Ndmsg) Serialize

func (msg *Ndmsg) Serialize() []byte

type Neigh

type Neigh struct {
	LinkIndex    int
	Family       int
	State        int
	Type         int
	Flags        int
	IP           net.IP
	HardwareAddr net.HardwareAddr
}

Neigh represents a link layer neighbor from netlink.

func NeighDeserialize

func NeighDeserialize(m []byte) (*Neigh, error)

func NeighList

func NeighList(linkIndex, family int) ([]Neigh, error)

NeighList gets a list of IP-MAC mappings in the system (ARP table). Equivalent to: `ip neighbor show`. The list can be filtered by link and ip family.

func (*Neigh) String

func (neigh *Neigh) String() string

String returns $ip/$hwaddr $label

type NsFd

type NsFd int

type NsPid

type NsPid int

type Protinfo

type Protinfo struct {
	Hairpin   bool
	Guard     bool
	FastLeave bool
	RootBlock bool
	Learning  bool
	Flood     bool
}

Protinfo represents bridge flags from netlink.

func LinkGetProtinfo

func LinkGetProtinfo(link Link) (Protinfo, error)

func (*Protinfo) String

func (prot *Protinfo) String() string

String returns a list of enabled flags

type Proto

type Proto uint8

Proto is an enum representing an ipsec protocol.

const (
	XFRM_PROTO_ROUTE2    Proto = syscall.IPPROTO_ROUTING
	XFRM_PROTO_ESP       Proto = syscall.IPPROTO_ESP
	XFRM_PROTO_AH        Proto = syscall.IPPROTO_AH
	XFRM_PROTO_HAO       Proto = syscall.IPPROTO_DSTOPTS
	XFRM_PROTO_COMP      Proto = syscall.IPPROTO_COMP
	XFRM_PROTO_IPSEC_ANY Proto = syscall.IPPROTO_RAW
)

func (Proto) String

func (p Proto) String() string

type Route

type Route struct {
	LinkIndex int
	Scope     Scope
	Dst       *net.IPNet
	Src       net.IP
	Gw        net.IP
}

Route represents a netlink route. A route is associated with a link, has a destination network, an optional source ip, and optional gateway. Advanced route parameters and non-main routing tables are currently not supported.

func RouteGet

func RouteGet(destination net.IP) ([]Route, error)

RouteGet gets a route to a specific destination from the host system. Equivalent to: 'ip route get'.

func RouteList

func RouteList(link Link, family int) ([]Route, error)

RouteList gets a list of routes in the system. Equivalent to: `ip route show`. The list can be filtered by link and ip family.

func (Route) String

func (r Route) String() string

type Scope

type Scope uint8

Scope is an enum representing a route scope.

const (
	SCOPE_UNIVERSE Scope = syscall.RT_SCOPE_UNIVERSE
	SCOPE_SITE     Scope = syscall.RT_SCOPE_SITE
	SCOPE_LINK     Scope = syscall.RT_SCOPE_LINK
	SCOPE_HOST     Scope = syscall.RT_SCOPE_HOST
	SCOPE_NOWHERE  Scope = syscall.RT_SCOPE_NOWHERE
)

type Veth

type Veth struct {
	LinkAttrs
	PeerName string // veth on create only
}

Veth devices must specify PeerName on create

func (*Veth) Attrs

func (veth *Veth) Attrs() *LinkAttrs

func (*Veth) Type

func (veth *Veth) Type() string

type Vlan

type Vlan struct {
	LinkAttrs
	VlanId int
}

Vlan links have ParentIndex set in their Attrs()

func (*Vlan) Attrs

func (vlan *Vlan) Attrs() *LinkAttrs

func (*Vlan) Type

func (vlan *Vlan) Type() string

type Vxlan

type Vxlan struct {
	LinkAttrs
	VxlanId      int
	VtepDevIndex int
	SrcAddr      net.IP
	Group        net.IP
	TTL          int
	TOS          int
	Learning     bool
	Proxy        bool
	RSC          bool
	L2miss       bool
	L3miss       bool
	NoAge        bool
	Age          int
	Limit        int
	Port         int
	PortLow      int
	PortHigh     int
}

func (*Vxlan) Attrs

func (vxlan *Vxlan) Attrs() *LinkAttrs

func (*Vxlan) Type

func (vxlan *Vxlan) Type() string

type XfrmPolicy

type XfrmPolicy struct {
	Dst      *net.IPNet
	Src      *net.IPNet
	Dir      Dir
	Priority int
	Index    int
	Tmpls    []XfrmPolicyTmpl
}

XfrmPolicy represents an ipsec policy. It represents the overlay network and has a list of XfrmPolicyTmpls representing the base addresses of the policy.

func XfrmPolicyList

func XfrmPolicyList(family int) ([]XfrmPolicy, error)

XfrmPolicyList gets a list of xfrm policies in the system. Equivalent to: `ip xfrm policy show`. The list can be filtered by ip family.

type XfrmPolicyTmpl

type XfrmPolicyTmpl struct {
	Dst   net.IP
	Src   net.IP
	Proto Proto
	Mode  Mode
	Reqid int
}

XfrmPolicyTmpl encapsulates a rule for the base addresses of an ipsec policy. These rules are matched with XfrmState to determine encryption and authentication algorithms.

type XfrmState

type XfrmState struct {
	Dst          net.IP
	Src          net.IP
	Proto        Proto
	Mode         Mode
	Spi          int
	Reqid        int
	ReplayWindow int
	Auth         *XfrmStateAlgo
	Crypt        *XfrmStateAlgo
	Encap        *XfrmStateEncap
}

XfrmState represents the state of an ipsec policy. It optionally contains an XfrmStateAlgo for encryption and one for authentication.

func XfrmStateList

func XfrmStateList(family int) ([]XfrmState, error)

XfrmStateList gets a list of xfrm states in the system. Equivalent to: `ip xfrm state show`. The list can be filtered by ip family.

type XfrmStateAlgo

type XfrmStateAlgo struct {
	Name        string
	Key         []byte
	TruncateLen int // Auth only
}

XfrmStateAlgo represents the algorithm to use for the ipsec encryption.

type XfrmStateEncap

type XfrmStateEncap struct {
	Type            EncapType
	SrcPort         int
	DstPort         int
	OriginalAddress net.IP
}

XfrmEncap represents the encapsulation to use for the ipsec encryption.

Directories

Path Synopsis
Package nl has low level primitives for making Netlink calls.
Package nl has low level primitives for making Netlink calls.

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL