After 8 months of development published release of the free hypervisor Xen 4.21. Companies such as Amazon, Arm, EPAM Systems and AMD took part in the development of the new release. The project code is written in C and distributed under the GPLv2+ license. Development is carried out as part of the Linux Foundation.
Key changes in Xen 4.21:
- Full support for the Linux stubdomain, which allows you to organize the execution of components for device emulation under a separate unprivileged user. The Linux stubomains model was developed by the QUBES OS project and supports the use of emulation drivers from recent releases of QEMU, as well as related features available in QEMU for guest systems.
- Changes for x86-based systems:
- Addedsupport for the new PDX structure compression algorithm (Page inDeX), which improves Xen performance on Intel Sierra Forest and Granite Rapids processors with non-linear physical memory mapping (sparse memory map), allowing the presence of empty areas.
- Added the ability to control the caching attribute in MTRR (Memory Type Range Registers) for the BAR (Base Address) register Register) is the underlying PCI device of the Xen platform used by guests in HVM mode. By default, the UC (uncacheable) attribute is set in MTRR for the reflected memory of PCI devices, which makes sense for real PCI devices, but only reduces the performance of the Xen platform PCI device.
- For AMD processors a new driver amd-cppc/amd-cppc-epp has been added, which controls the change in CPU frequency to achieve optimal performance. The driver uses the CPPC (Collaborative Processor Performance Control) mechanism, which, unlike acpi-cpufreq, is not limited to three performance levels and allows you to more quickly respond to state changes.
- In the version of the xenstore-stubdom layer, used to run xenstored without Dom0 privileges, for paravirtualized environments in PVH mode implemented live update support.
- For Dom0 in PVH implemented support for Resizable BAR technology (Resizable Base Address Register),
allowing you to immediately access the entire memory of a PCI Express device, and not just in relation to blocks. - The ability to forward PCI devices to domU environments in HVM (full hardware virtualization) mode has been implemented in configurations in which dom0 operates in PVH mode (a hybrid of hardware virtualization and paravirtualization).
/Reports, release notes, official announcements.