On 10/31/18 2:39 PM, Marc Orr wrote: > That makes sense. But my specific concern is the code I've copied > below, from arch/x86/kvm/x86.c. Notice on a system where > guest_fpu.state is a fregs_state, this code would generate garbage for > some fields. With the new code we're talking about, it will cause > memory corruption. But maybe it's not possible to run this code on a > system with an fregs_state, because such systems would predate VMX? Ahh, got it. So, you *can* clear X86_FEATURE_* bits from the kernel command-line, so it's theoretically possible to have a system that supports VMX, but doesn't support a modern MMU. It's obviously not well tested. :) The KVM code you pasted, to be "correct" should probably be checking X86_FEATURE_FXSR and X86_FEATURE_FPU *somewhere*.
