Okay, so I need to classify this bug report from QEMU into one of the given categories: mistranslation, assembly, other, device, graphic, socket, network, KVM, boot, vnc, debug, files, permissions, performance, kernel, peripherals, hypervisor. Let me read through the problem description carefully. The user is reporting that using `-M none` causes QEMU to fail on ARM64 platforms with a max IPA size less than 40 bits. They mention that KVM initialization fails because `virt_machine_class_init()` isn't called, which sets `mc->kvm_type`. This leads to the kernel defaulting to 40 bits, which isn't supported by some devices like Apple M1. The problem seems related to how QEMU initializes KVM on ARM64 when using `-M none`. The issue is that without calling `virt_machine_class_init()`, the necessary checks for IPA size aren't done. This affects systems where a larger IPA size (like 40 bits) isn't supported, causing failures. Looking at the steps to reproduce, they're comparing two commands: one with `-M none` and another with `-M virt`. The first doesn't set KVM parameters correctly, leading to failure on systems with limited IPA. This indicates a problem in how KVM is handled when the machine type is 'none'. Considering the categories, the issue directly relates to KVM setup and initialization. It's not about translations (mistranslation) or assembly code lowering. The specific mention of `KVM_CAP_ARM_VM_IPA_SIZE` points towards a KVM-related problem. So, the appropriate category here is definitely 'KVM'. This bug affects the proper configuration and initialization of KVM in QEMU when using certain machine types on ARM64 systems. The issue pertains to the initialization of KVM within QEMU, specifically when using `-M none` on ARM64 platforms. The problem arises because `virt_machine_class_init()` isn't called, leading to incorrect IPA size handling and subsequent KVM failures. This is a direct issue with how KVM is being set up. **KVM**