This article possibly contains original research. (January 2012) (Learn how and when to remove this template message)
This article uses HTML markup. (February 2019)
|Stable release||5.1.10 (15 June 2019 ) [±]|
|Preview release||5.2-rc5 (16 June 2019 ) [±]|
|Type||Application programming interface|
|License||GNU General Public License|
The Netlink socket family is a Linux kernel interface used for inter-process communication (IPC) between both the kernel and userspace processes, and between different userspace processes, in a way similar to the Unix domain sockets. Similarly to the Unix domain sockets, and unlike INET sockets, Netlink communication cannot traverse host boundaries. However, while the Unix domain sockets use the file system namespace, Netlink processes are usually addressed by process identifiers (PIDs). 
Netlink is designed and used for transferring miscellaneous networking information between the kernel space and userspace processes. Networking utilities, such as the iproute2 family and the utilities used for configuring mac80211-based wireless drivers, use Netlink to communicate with the Linux kernel from userspace. Netlink provides a standard socket-based interface for userspace processes, and a kernel-side API for internal use by kernel modules. Originally, Netlink used the AF_NETLINK socket family.
Netlink was created by Alexey Kuznetsov as a more flexible alternative to the sophisticated but awkward
ioctl communication method used for setting and getting external socket options. The Linux kernel continues to support
ioctl for backward compatibility. It can be used as follows:
error = ioctl(ip_socket, ioctl_type, &value_result);
Netlink was first provided in the 2.0 series of the Linux kernel, implemented as a character device. By 2013, this interface is obsolete, but still forms an ioctl communication method; compare the use of
rtnetlink. The Netlink socket interface appeared in 2.2 series of the Linux kernel.
Unlike BSD sockets using Internet protocols such as TCP, where the message headers are autogenerated, the Netlink message header (available as struct nlmsghdr) must be prepared by the caller. The Netlink socket generally works in a SOCK_RAW-like mode, even if SOCK_DGRAM was used to create it.
The data portion then contains a subsystem-specific message that may be further nested.
The AF_NETLINK family offers multiple protocol subsets. Each interfaces to a different kernel component and has a different messaging subset. The subset is referenced by the protocol field in the socket call:
int socket(AF_NETLINK, SOCK_DGRAM or SOCK_RAW, protocol)
Lacking a standard, SOCK_DGRAM and SOCK_RAW are not guaranteed to be implemented in a given Linux (or other OS) release. Some sources state that both options are legitimate, and the reference below from Red Hat states that SOCK_RAW is always the parameter. However, iproute2 uses both interchangeably.
A non-exhaustive list of the supported protocol entries follows:
NETLINK_ROUTE provides routing and link information. This information is used primarily for user-space routing daemons. Linux implements a large subset of messages:
- Link layer: RTM_NEWLINK, RTM_DELLINK, RTM_GETLINK, RTM_SETLINK
- Address settings: RTM_NEWADDR, RTM_DELADDR, RTM_GETADDR
- Routing tables: RTM_NEWROUTE, RTM_DELROUTE, RTM_GETROUTE
- Neighbor cache: RTM_NEWNEIGH, RTM_DELNEIGH, RTM_GETNEIGH
- Routing rules: RTM_NEWRULE, RTM_DELRULE, RTM_GETRULE
- Queuing discipline settings: RTM_NEWQDISC, RTM_DELQDISC, RTM_GETQDISC
- Traffic classes used with queues: RTM_NEWTCLASS, RTM_DELTCLASS, RTM_GETTCLASS
- Traffic filters: RTM_NEWTFILTER, RTM_DELTFILTER, RTM_GETTFILTER
- Others: RTM_NEWACTION, RTM_DELACTION, RTM_GETACTION, RTM_NEWPREFIX, RTM_GETPREFIX, RTM_GETMULTICAST, RTM_GETANYCAST, RTM_NEWNEIGHTBL, RTM_GETNEIGHTBL, RTM_SETNEIGHTBL
NETLINK_FIREWALL provides an interface for a user-space app to receive packets from the firewall.
NETLINK_ARPD provides an interface to manage the ARP table from user-space.
NETLINK_AUDIT provides an interface to the audit subsystem found in Linux kernel versions 2.6.6 and later.
NETLINK_IP6_FW provides an interface to transport packets from netfilter to user-space.
Users can add a Netlink handler in their own kernel routines. This allows the development of additional Netlink protocols to address new kernel modules.
- Comparison of open-source wireless drivers – mac80211-based drivers rely on netlink as the API to user space
- Kroah-Hartman, Greg (15 June 2019). "Linux 5.1.10". LKML (Mailing list). Retrieved 15 June 2019.
- Torvalds, Linus (16 June 2019). "Linux 5.2-rc5". LKML (Mailing list). Retrieved 16 June 2019.
- "NETLINK(7) - Linux Programmer's Manual".
- "kernel/git/torvalds/linux.git: root/net/core/rtnetlink.c". Linux kernel source tree. kernel.org. Retrieved 2014-05-27.
Crowcroft, Jon; Phillips, Iain, eds. (2002). TCP/IP and Linux protocol implementation: systems code for the Linux Internet. Wiley Networking Council series. Wiley. p. 624. ISBN 9780471408826. Retrieved 2013-05-21.
All rtnetlink messages consist of a netlink message header and appended attributes.
- Why and How to Use Netlink Sockets
- Pablo Neira Ayuso, Rafael M. Gasca, Laurent Lefèvre. Communicating between the kernel and user-space in Linux using Netlink sockets. Software: Practice and Experience, 40(9):797-810, August 2010
- Why and How to Use Netlink Sockets
- RFC 3549
- http://netfilter.org/projects/libmnl — “Minimalist Library for Netlink” — userspace library for construction and parsing of Netlink messages
- http://www.infradead.org/~tgr/libnl = "Netlink Protocol Library Suite" - full functional library covering almost all aspects of working with netlink sockets
- Manipulating the Networking Environment Using RTNETLINK
- Netlink Sockets – An Overview
- Netlink Protocol Library Suite
- "Linux Kernel Networking" by Rami Rosen, Apress 2013: Chapter 2, Netlink Sockets