npressfetimg-8300.png

Linux 5.19 KVM & Xen Changes Readied For This Next Kernel – Phoronix

Both the KVM and Xen changes are ready to go with the other code piling on for the Linux 5.19 merge window.

The KVM and Xen feature code for Linux 5.19 doesn’t have any major breakthrough features but a number of refinements and other incremental improvements being made for these important components to the open-source virtualization stack.

The Kernel-based Virtual Machine for x86/x86_64 adds new ioctls to set/get the TSC frequency for an entire VM, the ability for user-space to opt-out of hypercall patching, AMD Secure Encrypted Virtualization (SEV) improvements, and various AMD nested virtualization improvements.

KVM for RISC-V has Sv57x4 support for handling more virtual and physical memory. With Linux 5.18 RISC-V added Sv57 support and now it’s being built into the KVM side as well. The latest QEMU already has G-stage Sv57x4 mode support. RISC-V with this pull also has various other low-level updates.

KVM virtualization for Arm brings guard pages for EL2 stacks, support for the ARMv8.6 WFxT extension for WFIT/WFET, support for saving/restoring the set of hypercalls exposed to the guest, and various other fixes.

The complete set of KVM feature changes intended for Linux 5.19 is outlined via this pull.

Sent out separately this week were also the Xen updates for Linux 5.19. Xen continues seeing work to protect against possible malicious back-end driver code, support for Xen setups with multiple domains created at boot time, decoupling of the Xen SCSI para-virtualization (PV) interface from kernel internals, and other low-level code work. The Xen changes are outlined here.

Source: https://www.phoronix.com/scan.php?page=news_item&px=Linux-5.19-KVM-Xen

Related Posts