Re: [PATCH v14 8/9] KVM: x86: virtualize cpuid faulting

From: Andy Lutomirski
Date: Fri Jul 27 2018 - 16:46:14 EST




> On Jul 27, 2018, at 1:28 PM, Jim Mattson <jmattson@xxxxxxxxxx> wrote:
>
>> On Fri, Jul 27, 2018 at 12:41 PM, Andy Lutomirski <luto@xxxxxxxxxx> wrote:
>>> On Wed, Feb 8, 2017 at 12:09 AM, Kyle Huey <me@xxxxxxxxxxxx> wrote:
>>> Hardware support for faulting on the cpuid instruction is not required to
>>> emulate it, because cpuid triggers a VM exit anyways. KVM handles the relevant
>>> MSRs (MSR_PLATFORM_INFO and MSR_MISC_FEATURES_ENABLE) and upon a
>>> cpuid-induced VM exit checks the cpuid faulting state and the CPL.
>>> kvm_require_cpl is even kind enough to inject the GP fault for us.
>>>
>>> Signed-off-by: Kyle Huey <khuey@xxxxxxxxxxxx>
>>> Reviewed-by: David Matlack <dmatlack@xxxxxxxxxx>
>>> ---
>>> ...
>>> @@ -7613,16 +7636,19 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event)
>>>
>>> kvm_clear_async_pf_completion_queue(vcpu);
>>> kvm_async_pf_hash_reset(vcpu);
>>> vcpu->arch.apf.halted = false;
>>>
>>> if (!init_event) {
>>> kvm_pmu_reset(vcpu);
>>> vcpu->arch.smbase = 0x30000;
>>> +
>>> + vcpu->arch.msr_platform_info = MSR_PLATFORM_INFO_CPUID_FAULT;
>>> + vcpu->arch.msr_misc_features_enables = 0;
>>
>> Jim, I assume you're worried about this bit? It seems like
>> msr_platform_info should maybe be initialized to zero to avoid causing
>> an unintended migration issue.
>
> Initializing this bit to zero helps with migration, but then if the
> CPUID faulting bits in both MSRs are set, userspace has to take pains
> to ensure that MSR_PLATFORM_INFO is restored first, or the
> MSR_MISC_FEATURES_ENABLES value will be rejected.

The code could drop the constraint and just let the entry possibly fail if the MSRs are set wrong

>
> I'm also concerned about the 0 in the "Maximum Non-Turbo Ratio" field
> feeding into someone's calculated TSC frequency.

Hmm. I donât have a good answer to that. Are there any real CPUs that have this MSR but donât have that field?