Re: [PATCH v2 2/7] KVM: x86: Extract VMXON and EFER.SVME enablement to kernel

[Binbin Wu]

On 12/19/2025 11:40 PM, Sean Christopherson wrote:
> On Fri, Dec 19, 2025, Xu Yilun wrote:
>> On Wed, Dec 17, 2025 at 11:01:59AM -0800, Sean Christopherson wrote:
>>> On Wed, Dec 17, 2025, Xu Yilun wrote:
>>>> Is it better we explicitly assert the preemption for x86_virt_get_cpu()
>>>> rather than embed the check in __this_cpu_inc_return()? We are not just
>>>> protecting the racing for the reference counter. We should ensure the
>>>> "counter increase + x86_virt_call(get_cpu)" can't be preempted.
>>>
>>> I don't have a strong preference.  Using __this_cpu_inc_return() without any
>>> nearby preemption_{enable,disable}() calls makes it quite clears that preemption
>>> is expected to be disabled by the caller.  But I'm also ok being explicit.
>>
>> Looking into __this_cpu_inc_return(), it finally calls
>> check_preemption_disabled() which doesn't strictly requires preemption.
>> It only ensures the context doesn't switch to another CPU. If the caller
>> is in cpuhp context, preemption is possible.
> 
> Hmm, right, the cpuhp thread is is_percpu_thread(), and KVM's hooks aren't
> considered atomic and so run with IRQs enabled.  In practice, it's "fine", because
> TDX also exclusively does x86_virt_get_cpu() from cpuhp, i.e. the two users are
> mutually exclusive, but relying on that behavior is gross.
> 

Side topic:
Isn't the name of check_preemption_disabled() confusing and arguably misleading?