https://bugzilla.kernel.org/show_bug.cgi?id=198843 --- Comment #4 from David Hildenbrand (david@xxxxxxxxxx) --- Hi, (I'm also seeing what you describe under 2. on my TP and haven't found the root cause for it yet - well I haven't really started looking :) ). I assume the problem is not related. When the last guest is stopped, KVM tries to disable VMX. You seem to be getting a general protection fault trying to write to CR4 (native_write_cr4). static void kvm_cpu_vmxoff(void) { asm volatile (__ex(ASM_VMX_VMXOFF) : : : "cc"); intel_pt_handle_vmx(0); cr4_clear_bits(X86_CR4_VMXE); } So it looks like your hypervisor doesn't properly let you clear X86_CR4_VMXE - although you were allowed to enable it in hardware_enable(). At least that's what I see in this log. Maybe the google provided hypervisor has a BUG in that code? And older linux kernels maxbe disabled VMX when unloading the module instead, not when the las guest was released, that's why you see the error only with this kernel. -- You are receiving this mail because: You are watching the assignee of the bug.
