Yang Weijiang <[email protected]> writes:
> CET in nested guest over Hyper-V is not supported for now. Relevant
> enabling patches will be posted as a separate patch series.
>
> Suggested-by: Paolo Bonzini <[email protected]>
> Signed-off-by: Yang Weijiang <[email protected]>
> ---
> arch/x86/kvm/vmx/evmcs.h | 6 ++++--
> 1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/arch/x86/kvm/vmx/evmcs.h b/arch/x86/kvm/vmx/evmcs.h
> index bd41d9462355..25588694eb04 100644
> --- a/arch/x86/kvm/vmx/evmcs.h
> +++ b/arch/x86/kvm/vmx/evmcs.h
> @@ -59,8 +59,10 @@ DECLARE_STATIC_KEY_FALSE(enable_evmcs);
> SECONDARY_EXEC_SHADOW_VMCS | \
> SECONDARY_EXEC_TSC_SCALING | \
> SECONDARY_EXEC_PAUSE_LOOP_EXITING)
> -#define EVMCS1_UNSUPPORTED_VMEXIT_CTRL (VM_EXIT_LOAD_IA32_PERF_GLOBAL_CTRL)
> -#define EVMCS1_UNSUPPORTED_VMENTRY_CTRL (VM_ENTRY_LOAD_IA32_PERF_GLOBAL_CTRL)
> +#define EVMCS1_UNSUPPORTED_VMEXIT_CTRL (VM_EXIT_LOAD_IA32_PERF_GLOBAL_CTRL |
> \
> + VM_EXIT_LOAD_CET_STATE)
> +#define EVMCS1_UNSUPPORTED_VMENTRY_CTRL (VM_ENTRY_LOAD_IA32_PERF_GLOBAL_CTRL
> | \
> + VM_ENTRY_LOAD_CET_STATE)
> #define EVMCS1_UNSUPPORTED_VMFUNC (VMX_VMFUNC_EPTP_SWITCHING)
>
> #if IS_ENABLED(CONFIG_HYPERV)
This should be enough when we run KVM on Hyper-V using eVMCS, however,
it may not suffice when we run Hyper-V on KVM using eVMCS: there's still
no corresponding eVMCS fields so CET can't be used. In case Hyper-V is
smart enough it won't use the feature, however, it was proven to be 'not
very smart' in the past, see nested_evmcs_filter_control_msr(). I'm
wondering if we should also do
diff --git a/arch/x86/kvm/vmx/evmcs.c b/arch/x86/kvm/vmx/evmcs.c
index 41f24661af04..9f81db51fd8b 100644
--- a/arch/x86/kvm/vmx/evmcs.c
+++ b/arch/x86/kvm/vmx/evmcs.c
@@ -351,11 +351,11 @@ void nested_evmcs_filter_control_msr(u32 msr_index, u64
*pdata)
switch (msr_index) {
case MSR_IA32_VMX_EXIT_CTLS:
case MSR_IA32_VMX_TRUE_EXIT_CTLS:
- ctl_high &= ~VM_EXIT_LOAD_IA32_PERF_GLOBAL_CTRL;
+ ctl_high &= ~EVMCS1_UNSUPPORTED_VMEXIT_CTRL;
break;
case MSR_IA32_VMX_ENTRY_CTLS:
case MSR_IA32_VMX_TRUE_ENTRY_CTLS:
- ctl_high &= ~VM_ENTRY_LOAD_IA32_PERF_GLOBAL_CTRL;
+ ctl_high &= ~EVMCS1_UNSUPPORTED_VMENTRY_CTRL;
break;
case MSR_IA32_VMX_PROCBASED_CTLS2:
ctl_high &= ~SECONDARY_EXEC_VIRTUALIZE_APIC_ACCESSES;
to be on the safe side.
--
Vitaly
