1 .. SPDX-License-Identifier: BSD-3-Clause
2 Copyright(c) 2010-2015 Intel Corporation.
3 Copyright 2017 Mellanox Technologies, Ltd
6 .. _linux_gsg_linux_drivers:
11 Different PMDs may require different kernel drivers in order to work properly.
12 Depends on the PMD being used, a corresponding kernel driver should be load
13 and bind to the network ports.
18 A small kernel module to set up the device, map device memory to user-space and register interrupts.
19 In many cases, the standard ``uio_pci_generic`` module included in the Linux kernel
20 can provide the uio capability. This module can be loaded using the command:
22 .. code-block:: console
24 sudo modprobe uio_pci_generic
28 ``uio_pci_generic`` module doesn't support the creation of virtual functions.
30 As an alternative to the ``uio_pci_generic``, the DPDK also includes the igb_uio
31 module which can be found in the kmod subdirectory referred to above. It can
32 be loaded as shown below:
34 .. code-block:: console
37 sudo insmod kmod/igb_uio.ko
41 ``igb_uio`` module is disabled by default starting from ``DPDK v20.02``.
42 To build it, the config option ``CONFIG_RTE_EAL_IGB_UIO`` should be enabled.
43 It is planned to move ``igb_uio`` module to a different git repository.
47 For some devices which lack support for legacy interrupts, e.g. virtual function
48 (VF) devices, the ``igb_uio`` module may be needed in place of ``uio_pci_generic``.
52 If UEFI secure boot is enabled, the Linux kernel may disallow the use of
53 UIO on the system. Therefore, devices for use by DPDK should be bound to the
54 ``vfio-pci`` kernel module rather than ``igb_uio`` or ``uio_pci_generic``.
55 For more details see :ref:`linux_gsg_binding_kernel` below.
59 If the devices used for DPDK are bound to the ``uio_pci_generic`` kernel module,
60 please make sure that the IOMMU is disabled or passthrough. One can add
61 ``intel_iommu=off`` or ``amd_iommu=off`` or ``intel_iommu=on iommu=pt``in GRUB
62 command line on x86_64 systems, or add ``iommu.passthrough=1`` on arm64 system.
64 Since DPDK release 1.7 onward provides VFIO support, use of UIO is optional
65 for platforms that support using VFIO.
70 A more robust and secure driver in compare to the ``UIO``, relying on IOMMU protection.
71 To make use of VFIO, the ``vfio-pci`` module must be loaded:
73 .. code-block:: console
75 sudo modprobe vfio-pci
77 Note that in order to use VFIO, your kernel must support it.
78 VFIO kernel modules have been included in the Linux kernel since version 3.6.0 and are usually present by default,
79 however please consult your distributions documentation to make sure that is the case.
81 Also, to use VFIO, both kernel and BIOS must support and be configured to use IO virtualization (such as IntelĀ® VT-d).
85 ``vfio-pci`` module doesn't support the creation of virtual functions.
87 For proper operation of VFIO when running DPDK applications as a non-privileged user, correct permissions should also be set up.
88 This can be done by using the DPDK setup script (called dpdk-setup.sh and located in the usertools directory).
92 VFIO can be used without IOMMU. While this is just as unsafe as using UIO, it does make it possible for the user to keep the degree of device access and programming that VFIO has, in situations where IOMMU is not available.
94 .. _bifurcated_driver:
99 PMDs which use the bifurcated driver co-exists with the device kernel driver.
100 On such model the NIC is controlled by the kernel, while the data
101 path is performed by the PMD directly on top of the device.
103 Such model has the following benefits:
105 - It is secure and robust, as the memory management and isolation
106 is done by the kernel.
107 - It enables the user to use legacy linux tools such as ``ethtool`` or
108 ``ifconfig`` while running DPDK application on the same network ports.
109 - It enables the DPDK application to filter only part of the traffic,
110 while the rest will be directed and handled by the kernel driver.
111 The flow bifurcation is performed by the NIC hardware.
112 As an example, using :ref:`flow_isolated_mode` allows to choose
113 strictly what is received in DPDK.
115 More about the bifurcated driver can be found in
116 `Mellanox Bifurcated DPDK PMD
117 <https://www.dpdk.org/wp-content/uploads/sites/35/2016/10/Day02-Session04-RonyEfraim-Userspace2016.pdf>`__.
119 .. _linux_gsg_binding_kernel:
121 Binding and Unbinding Network Ports to/from the Kernel Modules
122 --------------------------------------------------------------
126 PMDs Which use the bifurcated driver should not be unbind from their kernel drivers. this section is for PMDs which use the UIO or VFIO drivers.
128 As of release 1.4, DPDK applications no longer automatically unbind all supported network ports from the kernel driver in use.
129 Instead, in case the PMD being used use the UIO or VFIO drivers, all ports that are to be used by an DPDK application must be bound to the
130 ``uio_pci_generic``, ``igb_uio`` or ``vfio-pci`` module before the application is run.
131 For such PMDs, any network ports under Linux* control will be ignored and cannot be used by the application.
133 To bind ports to the ``uio_pci_generic``, ``igb_uio`` or ``vfio-pci`` module for DPDK use,
134 and then subsequently return ports to Linux* control,
135 a utility script called dpdk-devbind.py is provided in the usertools subdirectory.
136 This utility can be used to provide a view of the current state of the network ports on the system,
137 and to bind and unbind those ports from the different kernel modules, including the uio and vfio modules.
138 The following are some examples of how the script can be used.
139 A full description of the script and its parameters can be obtained by calling the script with the ``--help`` or ``--usage`` options.
140 Note that the uio or vfio kernel modules to be used, should be loaded into the kernel before
141 running the ``dpdk-devbind.py`` script.
145 Due to the way VFIO works, there are certain limitations to which devices can be used with VFIO.
146 Mainly it comes down to how IOMMU groups work.
147 Any Virtual Function device can be used with VFIO on its own, but physical devices will require either all ports bound to VFIO,
148 or some of them bound to VFIO while others not being bound to anything at all.
150 If your device is behind a PCI-to-PCI bridge, the bridge will then be part of the IOMMU group in which your device is in.
151 Therefore, the bridge driver should also be unbound from the bridge PCI device for VFIO to work with devices behind the bridge.
155 While any user can run the dpdk-devbind.py script to view the status of the network ports,
156 binding or unbinding network ports requires root privileges.
158 To see the status of all network ports on the system:
160 .. code-block:: console
162 ./usertools/dpdk-devbind.py --status
164 Network devices using DPDK-compatible driver
165 ============================================
166 0000:82:00.0 '82599EB 10-GbE NIC' drv=uio_pci_generic unused=ixgbe
167 0000:82:00.1 '82599EB 10-GbE NIC' drv=uio_pci_generic unused=ixgbe
169 Network devices using kernel driver
170 ===================================
171 0000:04:00.0 'I350 1-GbE NIC' if=em0 drv=igb unused=uio_pci_generic *Active*
172 0000:04:00.1 'I350 1-GbE NIC' if=eth1 drv=igb unused=uio_pci_generic
173 0000:04:00.2 'I350 1-GbE NIC' if=eth2 drv=igb unused=uio_pci_generic
174 0000:04:00.3 'I350 1-GbE NIC' if=eth3 drv=igb unused=uio_pci_generic
176 Other network devices
177 =====================
180 To bind device ``eth1``,``04:00.1``, to the ``uio_pci_generic`` driver:
182 .. code-block:: console
184 ./usertools/dpdk-devbind.py --bind=uio_pci_generic 04:00.1
188 .. code-block:: console
190 ./usertools/dpdk-devbind.py --bind=uio_pci_generic eth1
192 To restore device ``82:00.0`` to its original kernel binding:
194 .. code-block:: console
196 ./usertools/dpdk-devbind.py --bind=ixgbe 82:00.0