Microsoft wants to create a complete virtualization stack with Linux and Hyper-V
Microsoft engineer Wei Liu recently submitted a series of kernel patches for Hyper-V code for Linux. The purpose is to make Linux run as a root partition on Microsoft Hypervisor (Hyper-V).
Hyper-V is a very well-known virtualization technology from Microsoft that provides hardware virtualization. Not only allows you to create virtual hard disks and network switches, but you can also run multiple virtual operating systems on the Windows platform.
The root partition in the Microsoft Hypervisor organization hierarchy is similar to Xeon’s Dom0, which is used to sequentially start and manage non-privileged domains. According to Microsoft official documentation, the root partition can directly access physical I/O devices, and the virtualization stack in it can provide memory management for virtual machines, management APIs, and virtualized I/O devices. It also implements device emulation, provides integrated electronic component (IDE) disk controller and PS/2 input, and supports Hyper-C specific device synthesis to improve performance and reduce overhead.
As Phoronix said, nowadays, although Linux has provided support for Hyper-V, and a half or more of the virtual machines on the Microsoft Azure cloud platform is based on Linux. However, the current mainline Linux kernel still does not allow it to be used for the root partition of Hyper-V. Therefore, Microsoft now hopes to create a complete virtualization stack using Linux by introducing Linux root partition support for its Hyper-V.
The content of the email pointed out that the patch is a basic change to the Linux Hyper-V code. Microsoft hopes to use Linux and Hyper-V to create a complete virtualization stack. There will be a series of patches that will provide a device node (/dev/mshv) so that userspace can create and run virtual machines. At present, the team has completed the porting of Cloud Hypervisor, and since late July, it has been possible to start Linux guests with Virtio devices.
