8 PCI(e) passthrough is a mechanism to give a virtual machine control over
9 a PCI device from the host. This can have some advantages over using
10 virtualized hardware, for example lower latency, higher performance, or more
11 features (e.g., offloading).
13 But, if you pass through a device to a virtual machine, you cannot use that
14 device anymore on the host or in any other VM.
19 Since passthrough is a feature which also needs hardware support, there are
20 some requirements to check and preparations to be done to make it work.
25 Your hardware needs to support `IOMMU` (*I*/*O* **M**emory **M**anagement
26 **U**nit) interrupt remapping, this includes the CPU and the mainboard.
28 Generally, Intel systems with VT-d, and AMD systems with AMD-Vi support this.
29 But it is not guaranteed that everything will work out of the box, due
30 to bad hardware implementation and missing or low quality drivers.
32 Further, server grade hardware has often better support than consumer grade
33 hardware, but even then, many modern system can support this.
35 Please refer to your hardware vendor to check if they support this feature
36 under Linux for your specific setup.
42 Once you ensured that your hardware supports passthrough, you will need to do
43 some configuration to enable PCI(e) passthrough.
48 First, the IOMMU support has to be enabled in your BIOS/UEFI. Most often, that
49 options is named `IOMMU` or `VT-d`, but check the manual for your motherboard
50 for the exact option you need to enable.
52 Then, the IOMMU has to be activated on the
53 xref:sysboot_edit_kernel_cmdline[kernel commandline].
55 The command line parameters are:
62 * for AMD CPUs it should be enabled automatically.
66 You have to make sure the following modules are loaded. This can be achieved by
67 adding them to `'/etc/modules''
76 [[qm_pci_passthrough_update_initramfs]]
77 After changing anything modules related, you need to refresh your
78 `initramfs`. On {pve} this can be done by executing:
81 # update-initramfs -u -k all
86 Finally reboot to bring the changes into effect and check that it is indeed
90 # dmesg | grep -e DMAR -e IOMMU -e AMD-Vi
93 should display that `IOMMU`, `Directed I/O` or `Interrupt Remapping` is
94 enabled, depending on hardware and kernel the exact message can vary.
96 It is also important that the device(s) you want to pass through
97 are in a *separate* `IOMMU` group. This can be checked with:
100 # find /sys/kernel/iommu_groups/ -type l
103 It is okay if the device is in an `IOMMU` group together with its functions
104 (e.g. a GPU with the HDMI Audio device) or with its root port or PCI(e) bridge.
109 Some platforms handle their physical PCI(e) slots differently. So, sometimes
110 it can help to put the card in a another PCI(e) slot, if you do not get the
111 desired `IOMMU` group separation.
117 For some platforms, it may be necessary to allow unsafe interrupts.
118 For this add the following line in a file ending with `.conf' file in
122 options vfio_iommu_type1 allow_unsafe_interrupts=1
125 Please be aware that this option can make your system unstable.
128 GPU Passthrough Notes
129 ^^^^^^^^^^^^^^^^^^^^^
131 It is not possible to display the frame buffer of the GPU via NoVNC or SPICE on
132 the {pve} web interface.
134 When passing through a whole GPU or a vGPU and graphic output is wanted, one
135 has to either physically connect a monitor to the card, or configure a remote
136 desktop software (for example, VNC or RDP) inside the guest.
138 If you want to use the GPU as a hardware accelerator, for example, for
139 programs using OpenCL or CUDA, this is not required.
141 Host Device Passthrough
142 ~~~~~~~~~~~~~~~~~~~~~~~
144 The most used variant of PCI(e) passthrough is to pass through a whole
145 PCI(e) card, for example a GPU or a network card.
151 In this case, the host must not use the card. There are two methods to achieve
154 * pass the device IDs to the options of the 'vfio-pci' modules by adding
157 options vfio-pci ids=1234:5678,4321:8765
160 to a .conf file in */etc/modprobe.d/* where `1234:5678` and `4321:8765` are
161 the vendor and device IDs obtained by:
167 * blacklist the driver completely on the host, ensuring that it is free to bind
168 for passthrough, with
174 in a .conf file in */etc/modprobe.d/*.
176 For both methods you need to
177 xref:qm_pci_passthrough_update_initramfs[update the `initramfs`] again and
180 .Verify Configuration
182 To check if your changes were successful, you can use
188 and check your device entry. If it says
191 Kernel driver in use: vfio-pci
194 or the 'in use' line is missing entirely, the device is ready to be used for
197 [[qm_pci_passthrough_vm_config]]
200 To pass through the device you need to set the *hostpciX* option in the VM
201 configuration, for example by executing:
204 # qm set VMID -hostpci0 00:02.0
207 If your device has multiple functions (e.g., ``00:02.0`' and ``00:02.1`' ),
208 you can pass them through all together with the shortened syntax ``00:02`'.
209 This is equivalent with checking the ``All Functions`' checkbox in the
212 There are some options to which may be necessary, depending on the device
215 * *x-vga=on|off* marks the PCI(e) device as the primary GPU of the VM.
216 With this enabled the *vga* configuration option will be ignored.
218 * *pcie=on|off* tells {pve} to use a PCIe or PCI port. Some guests/device
219 combination require PCIe rather than PCI. PCIe is only available for 'q35'
222 * *rombar=on|off* makes the firmware ROM visible for the guest. Default is on.
223 Some PCI(e) devices need this disabled.
225 * *romfile=<path>*, is an optional path to a ROM file for the device to use.
226 This is a relative path under */usr/share/kvm/*.
230 An example of PCIe passthrough with a GPU set to primary:
233 # qm set VMID -hostpci0 02:00,pcie=on,x-vga=on
240 When passing through a GPU, the best compatibility is reached when using
241 'q35' as machine type, 'OVMF' ('EFI' for VMs) instead of SeaBIOS and PCIe
242 instead of PCI. Note that if you want to use 'OVMF' for GPU passthrough, the
243 GPU needs to have an EFI capable ROM, otherwise use SeaBIOS instead.
248 Another variant for passing through PCI(e) devices, is to use the hardware
249 virtualization features of your devices, if available.
251 'SR-IOV' (**S**ingle-**R**oot **I**nput/**O**utput **V**irtualization) enables
252 a single device to provide multiple 'VF' (**V**irtual **F**unctions) to the
253 system. Each of those 'VF' can be used in a different VM, with full hardware
254 features and also better performance and lower latency than software
257 Currently, the most common use case for this are NICs (**N**etwork
258 **I**nterface **C**ard) with SR-IOV support, which can provide multiple VFs per
259 physical port. This allows using features such as checksum offloading, etc. to
260 be used inside a VM, reducing the (host) CPU overhead.
266 Generally, there are two methods for enabling virtual functions on a device.
268 * sometimes there is an option for the driver module e.g. for some
275 which could be put file with '.conf' ending under */etc/modprobe.d/*.
276 (Do not forget to update your initramfs after that)
278 Please refer to your driver module documentation for the exact
279 parameters and options.
281 * The second, more generic, approach is using the `sysfs`.
282 If a device and driver supports this you can change the number of VFs on
283 the fly. For example, to setup 4 VFs on device 0000:01:00.0 execute:
286 # echo 4 > /sys/bus/pci/devices/0000:01:00.0/sriov_numvfs
289 To make this change persistent you can use the `sysfsutils` Debian package.
290 After installation configure it via */etc/sysfs.conf* or a `FILE.conf' in
296 After creating VFs, you should see them as separate PCI(e) devices when
297 outputting them with `lspci`. Get their ID and pass them through like a
298 xref:qm_pci_passthrough_vm_config[normal PCI(e) device].
303 For this feature, platform support is especially important. It may be necessary
304 to enable this feature in the BIOS/EFI first, or to use a specific PCI(e) port
305 for it to work. In doubt, consult the manual of the platform or contact its
308 Mediated Devices (vGPU, GVT-g)
309 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
311 Mediated devices are another method to reuse features and performance from
312 physical hardware for virtualized hardware. These are found most common in
313 virtualized GPU setups such as Intel's GVT-g and NVIDIA's vGPUs used in their
316 With this, a physical Card is able to create virtual cards, similar to SR-IOV.
317 The difference is that mediated devices do not appear as PCI(e) devices in the
318 host, and are such only suited for using in virtual machines.
324 In general your card's driver must support that feature, otherwise it will
325 not work. So please refer to your vendor for compatible drivers and how to
328 Intel's drivers for GVT-g are integrated in the Kernel and should work
329 with 5th, 6th and 7th generation Intel Core Processors, as well as E3 v4, E3
330 v5 and E3 v6 Xeon Processors.
332 To enable it for Intel Graphics, you have to make sure to load the module
333 'kvmgt' (for example via `/etc/modules`) and to enable it on the
334 xref:sysboot_edit_kernel_cmdline[Kernel commandline] and add the following parameter:
340 After that remember to
341 xref:qm_pci_passthrough_update_initramfs[update the `initramfs`],
342 and reboot your host.
347 To use a mediated device, simply specify the `mdev` property on a `hostpciX`
348 VM configuration option.
350 You can get the supported devices via the 'sysfs'. For example, to list the
351 supported types for the device '0000:00:02.0' you would simply execute:
354 # ls /sys/bus/pci/devices/0000:00:02.0/mdev_supported_types
357 Each entry is a directory which contains the following important files:
359 * 'available_instances' contains the amount of still available instances of
360 this type, each 'mdev' use in a VM reduces this.
361 * 'description' contains a short description about the capabilities of the type
362 * 'create' is the endpoint to create such a device, {pve} does this
363 automatically for you, if a 'hostpciX' option with `mdev` is configured.
365 Example configuration with an `Intel GVT-g vGPU` (`Intel Skylake 6700k`):
368 # qm set VMID -hostpci0 00:02.0,mdev=i915-GVTg_V5_4
371 With this set, {pve} automatically creates such a device on VM start, and
372 cleans it up again when the VM stops.
379 * link:/wiki/Pci_passthrough[PCI Passthrough Examples]