Re: [PATCH v19 022/130] KVM: x86/vmx: Refactor KVM VMX module init/exit functions

[Yin Fengwei]

On 3/12/24 10:15, Isaku Yamahata wrote:
>>> -
>>> -	__vmx_exit();
>>> -}
>>> -module_exit(vmx_exit);
>>> -
>>> -static int __init vmx_init(void)
>>> +int __init vmx_init(void)
>>>   {
>>>   	int r, cpu;
>>> -	if (!kvm_is_vmx_supported())
>>> -		return -EOPNOTSUPP;
>>> -
>>> -	/*
>>> -	 * Note, hv_init_evmcs() touches only VMX knobs, i.e. there's nothing
>>> -	 * to unwind if a later step fails.
>>> -	 */
>>> -	hv_init_evmcs();
>>> -
>>> -	/* vmx_hardware_disable() accesses loaded_vmcss_on_cpu. */
>>> -	for_each_possible_cpu(cpu)
>>> -		INIT_LIST_HEAD(&per_cpu(loaded_vmcss_on_cpu, cpu));
>>> -
>>> -	r = kvm_x86_vendor_init(&vt_init_ops);
>>> -	if (r)
>>> -		return r;
>>> -
>>>   	/*
>>>   	 * Must be called after common x86 init so enable_ept is properly set
>>>   	 * up. Hand the parameter mitigation value in which was stored in
>> I am wondering whether the first sentence of above comment should be
>> moved to vt_init()? So vt_init() has whole information about the init
>> sequence.
> If we do so, we should move the call of "vmx_setup_l1d_flush() to vt_init().
> I hesitated to remove static of vmx_setup_l1d_flush().
I meant this one:
 "Must be called after common x86 init so enable_ept is properly set up"

Not necessary to move vmx_setup_l1d_flush().

Regards
Yin, Fengwei


> --