Documentation/networking/vrf.rst
Source file repositories/reference/linux-study-clean/Documentation/networking/vrf.rst
File Facts
- System
- Linux kernel
- Corpus path
Documentation/networking/vrf.rst- Extension
.rst- Size
- 16853 bytes
- Lines
- 465
- Domain
- Support Tooling And Documentation
- Bucket
- Documentation
- Inferred role
- Support Tooling And Documentation: documentation
- Status
- atlas-only
Why This File Exists
Repository support layer: documentation, build tooling, samples, user-space helper tools, generated initramfs support, licenses, and validation utilities.
- Repository support layer: documentation, build tooling, samples, user-space helper tools, generated initramfs support, licenses, and validation utilities.
Dependency Surface
- No C-style include directives detected by the generator.
Detected Declarations
- No top-level syscall, struct, function, initcall, or export declaration detected by the generator.
Annotated Snippet
.. SPDX-License-Identifier: GPL-2.0
====================================
Virtual Routing and Forwarding (VRF)
====================================
The VRF Device
==============
The VRF device combined with ip rules provides the ability to create virtual
routing and forwarding domains (aka VRFs, VRF-lite to be specific) in the
Linux network stack. One use case is the multi-tenancy problem where each
tenant has their own unique routing tables and in the very least need
different default gateways.
Processes can be "VRF aware" by binding a socket to the VRF device. Packets
through the socket then use the routing table associated with the VRF
device. An important feature of the VRF device implementation is that it
impacts only Layer 3 and above so L2 tools (e.g., LLDP) are not affected
(ie., they do not need to be run in each VRF). The design also allows
the use of higher priority ip rules (Policy Based Routing, PBR) to take
precedence over the VRF device rules directing specific traffic as desired.
In addition, VRF devices allow VRFs to be nested within namespaces. For
example network namespaces provide separation of network interfaces at the
device layer, VLANs on the interfaces within a namespace provide L2 separation
and then VRF devices provide L3 separation.
Design
------
A VRF device is created with an associated route table. Network interfaces
are then enslaved to a VRF device::
+-----------------------------+
| vrf-blue | ===> route table 10
+-----------------------------+
| | |
+------+ +------+ +-------------+
| eth1 | | eth2 | ... | bond1 |
+------+ +------+ +-------------+
| |
+------+ +------+
| eth8 | | eth9 |
+------+ +------+
Packets received on an enslaved device and are switched to the VRF device
in the IPv4 and IPv6 processing stacks giving the impression that packets
flow through the VRF device. Similarly on egress routing rules are used to
send packets to the VRF device driver before getting sent out the actual
interface. This allows tcpdump on a VRF device to capture all packets into
and out of the VRF as a whole\ [1]_. Similarly, netfilter\ [2]_ and tc rules
can be applied using the VRF device to specify rules that apply to the VRF
domain as a whole.
.. [1] Packets in the forwarded state do not flow through the device, so those
packets are not seen by tcpdump. Will revisit this limitation in a
future release.
.. [2] Iptables on ingress supports PREROUTING with skb->dev set to the real
ingress device and both INPUT and PREROUTING rules with skb->dev set to
the VRF device. For egress POSTROUTING and OUTPUT rules can be written
using either the VRF device or real egress device.
Setup
-----
1. VRF device is created with an association to a FIB table.
e.g,::
ip link add vrf-blue type vrf table 10
ip link set dev vrf-blue up
Annotation
- Atlas domain: Support Tooling And Documentation / Documentation.
- Implementation status: atlas-only.
Implementation Notes
- This generated page is the file-by-file coverage layer; curated subsystem chapters should link here when they synthesize a multi-file control flow.
- Core OS pages should be promoted from atlas-only to deep-reviewed when they explain data structures, invariants, locking, lifecycle, and C implementation snippets.
- Driver-family pages are intentionally pattern-oriented unless they are part of the selected PCIe/NVMe representative device path.