Xen hypervisor port to RISC-V moving – slowly, but moving

Far from ready, but developer thinks it's helping the whole codebase

If RISC-V is to become a viable processor architecture for servers, its software ecosystem will need a solid hypervisor.

Which is why The Register noted in 2021 that a project had commenced to bring the open source Xen hypervisor to RISC-V, the open instruction set architecture.

And in 2023 we can report that effort is moving … slowly.

A February 7 post from Oleksii Kurochko, a hypervisor and kernel software engineer at Vates – the organization that created the XCP-ng cut of Xen – explains that the effort has taken a few steps forward and a few steps back.

Kurochko explained that some recent patches to Xen were promising – but also too complex to integrate easily.

"It was decided to start committing and reworking some patches step by step," he wrote.

To date, a patch allowing minimal Xen implementations on RISC-V has been merged, as has a cross-build test.

Also on Kurochko's to-do list:

• Introducing a smoke test and job to verify early printk functionality;
• Basic trap handling;
• Updating the smoke test to verify the functionality from asm/bug.h;
• Making some elements of Xen generic across x86, Arm and RISC-V;
• Basic MMU support (which he thinks is the next item he'll address);
• Various other bits of functionality, one at a time – smp support, console, enable xen/common, etc.

Kurochko excitedly added that developing Xen for RISC-V is non-trivial, but rated his work "a great demonstration that porting Xen to RISC-V could also help to make the Xen code base better and more generic. Even if it's more work in the short term, it's better for the project on the long run!"

"Work on porting is still in progress (I'm working full time on it), and it's far from being done," he concluded. "However, there's a lot of momentum going on with our work on RISC-V+Xen, as there is significant support and interest from the Xen community itself and even beyond!" ®