Use of Hypervisors - 2025.1 English - UG1304

AMD distributes a port for the Xen open source hypervisor for the Versal device. The Xen hypervisor provides the ability to run multiple operating systems on the same computing platform. Xen hypervisor, which runs directly on the hardware, is responsible for managing the CPU, memory, and interrupts. You can run multiple operating systems on the hypervisor, which are called domains, virtual machines (VMs), or a guest OS.

The Xen hypervisor provides the ability to concurrently run multiple operating systems and their standard applications with relative ease. It also makes it possible to give a guest OS direct access to specific peripherals. However, Xen does not provide a generic method to do so; peripheral-specific configurations are required.

Dom0 or control domain has special privileges including:

• Capability to access the hardware directly.
• Ability to handle access to the I/O functions of the system.
• Interaction with other virtual machines.
• Dynamic node programming to assign/remove the FPGA and generic device tree nodes to/from different running guests using a device tree binary overlay.
• Emulated TPM to provide each guest with a unique TPM for enabling secure and measured boot.

Each guest domain runs its own OS and application. Guest domains are completely isolated from the hardware.

For more information the Xen hypervisor, refer to the Xen Hypervisor page the AMD Wiki.

Xen hypervisor is available as a selectable component within the Yocto tools; alternatively, you can download the Xen hypervisor. With Linux and Xen software that is made available by AMD, it is possible to build custom Linux guest configurations. Guest OS other than Linux require additional software and effort from third-parties. For more information, see the PetaLinux Tools page.

In addition to the Xen hypervisor, other hypervisors are available through various AMD ecosystem partners. Visit the Embedded Software page for further details.