Re: [PATCH v2 05/18] x86/reboot: KVM: Disable SVM during reboot via virt/KVM reboot callback

["Huang, Kai" <kai.huang@intel.com>]

On Fri, 2023-03-10 at 13:42 -0800, Sean Christopherson wrote:
> Use the virt callback to disable SVM (and set GIF=1) during an emergency
> instead of blindly attempting to disable SVM.  Like the VMX case, if KVM
> (or an out-of-tree hypervisor) isn't loaded/active, SVM can't be in use.
> 
> Signed-off-by: Sean Christopherson <seanjc@google.com>

[...]

> -#if IS_ENABLED(CONFIG_KVM_INTEL)
> +#if IS_ENABLED(CONFIG_KVM_INTEL) || IS_ENABLED(CONFIG_KVM_AMD)
>  /* RCU-protected callback to disable virtualization prior to reboot. */
>  static cpu_emergency_virt_cb __rcu *cpu_emergency_virt_callback;
>  
> @@ -821,7 +821,7 @@ int crashing_cpu = -1;
>   */
>  void cpu_emergency_disable_virtualization(void)
>  {
> -#if IS_ENABLED(CONFIG_KVM_INTEL)
> +#if IS_ENABLED(CONFIG_KVM_INTEL) || IS_ENABLED(CONFIG_KVM_AMD)
>  	cpu_emergency_virt_cb *callback;
>  
>  	rcu_read_lock();
> @@ -830,8 +830,6 @@ void cpu_emergency_disable_virtualization(void)
>  		callback();
>  	rcu_read_unlock();
>  #endif
> -	/* KVM_AMD doesn't yet utilize the common callback. */
> -	cpu_emergency_svm_disable();
>  }

Shouldn't the callback be always present since you want to consider 'out-of-
tree' hypervisor case?