The release of QEMU 6.2 has been announced. As an emulator, QEMU allows you to run programs compiled for one hardware platform on a system with a completely different architecture, for example, running an ARM application on an x86-compatible PC. In virtualization mode, the performance of code execution in an isolated environment is close to that of a physical system due to direct execution of instructions on the CPU and the involvement of the Xen hypervisor or KVM module.
The project was originally created by Fabrice Bellard with the aim of enabling the execution of Linux executables compiled for the x86 platform on architectures other than x86. Over the years of development, support for full emulation for 14 hardware architectures has been added, and the number of emulated hardware devices has exceeded 400. More than 2300 changes from 189 developers have been made in preparation for version 6.2.
Key improvements added in QEMU 6.2:
- In the virtio-mem mechanism, which allows for hot-plugging and unplugging of memory to virtual machines, complete support for guest system memory dumps, pre-copy/post-copy migration operations, and the creation of background snapshots of the guest system has been added.
- Error handling for DEVICE_UNPLUG_GUEST_ERROR has been implemented in QMP (QEMU Machine Protocol), which occurs on the guest system side in case of failures during hardware hot-plug operations.
- The syntax of boot arguments processed in plugins for the classic TCG (Tiny Code Generator) has been expanded. The ‘cache’ plugin now supports multi-core systems.
- Support for the Intel Snowridge-v4 CPU model has been implemented in the x86 architecture emulator. Access to Intel SGX (Software Guard eXtensions) enclaves from guest systems has been added, using the /dev/sgx_vepc device on the host system side and the ‘memory-backend-epc’ backend in QEMU. For guest systems protected by AMD SEV (Secure Encrypted Virtualization), the capability for verifiable direct (bootloader-free) kernel launch has been added (enabled by setting the ‘kernel-hashes=on’ parameter in ‘sev-guest’).
- The ARM emulator on host systems with Apple Silicon chips has implemented support for the hardware acceleration mechanism "hvf" when launching guest systems based on the AArch64 architecture. Support for emulating Fujitsu A64FX processor models has been added. A new type of emulated machine "kudo-mbc" has been implemented. For 'virt' machines, support for emulating the ITS (Interrupt Translation Service) has been added, along with the capability to use more than 123 CPUs in emulation mode. For the emulated machines "xlnx-zcu102" and "xlnx-versal-virt," support for BBRAM and eFUSE devices has been added. For systems based on the Cortex-M55 chip, support for the mobile profile of MVE processor extensions is provided.
- The PowerPC architecture emulator has added initial support for the POWER10 DD2.0 CPU model. For the emulated "powernv" machines, enhanced support for the POWER10 architecture has been implemented, and for 'pseries' machines, FORM2 PAPR NUMA descriptions have been added.
- The RISC-V architecture emulator has added support for the instruction set extensions Zb[abcs]. For all emulated machines, the use of "host-user" and "numa mem" options has been permitted. Support for SiFive PWM (Pulse-width modulator) has been introduced.
- In the 68k architecture emulator, support for the NuBus bus proposed by Apple has been improved, including the ability to load ROM images and support for interrupt slots.
- In the qemu-nbd block device, the write operation caching mode ("writeback" instead of "writethrough") is enabled by default to match the behavior of qemu-img. An option "--selinux-label" has been added for setting SELinux labels on Unix sockets.
Source: opennet.ru
