Re: [PATCH] KVM: x86: use CPUID to locate host page table reserved bits

From: Sean Christopherson
Date: Wed Dec 04 2019 - 10:29:45 EST


On Wed, Dec 04, 2019 at 03:51:00PM +0100, Paolo Bonzini wrote:
> The comment in kvm_get_shadow_phys_bits refers to MKTME, but the same is actually
> true of SME and SEV. Just use CPUID[0x8000_0008].EAX[7:0] unconditionally, it is
> simplest and works even if memory is not encrypted.
>
> Cc: stable@xxxxxxxxxxxxxxx
> Reported-by: Tom Lendacky <thomas.lendacky@xxxxxxx>
> Signed-off-by: Paolo Bonzini <pbonzini@xxxxxxxxxx>
> ---
> arch/x86/kvm/mmu/mmu.c | 12 ++++--------
> 1 file changed, 4 insertions(+), 8 deletions(-)
>
> diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
> index 6f92b40d798c..8b8edfbdbaef 100644
> --- a/arch/x86/kvm/mmu/mmu.c
> +++ b/arch/x86/kvm/mmu/mmu.c
> @@ -538,15 +538,11 @@ void kvm_mmu_set_mask_ptes(u64 user_mask, u64 accessed_mask,
> static u8 kvm_get_shadow_phys_bits(void)
> {
> /*
> - * boot_cpu_data.x86_phys_bits is reduced when MKTME is detected
> - * in CPU detection code, but MKTME treats those reduced bits as
> - * 'keyID' thus they are not reserved bits. Therefore for MKTME
> - * we should still return physical address bits reported by CPUID.
> + * boot_cpu_data.x86_phys_bits is reduced when MKTME or SME are detected
> + * in CPU detection code, but the processor treats those reduced bits as
> + * 'keyID' thus they are not reserved bits. Therefore KVM needs to look at
> + * the physical address bits reported by CPUID.
> */
> - if (!boot_cpu_has(X86_FEATURE_TME) ||
> - WARN_ON_ONCE(boot_cpu_data.extended_cpuid_level < 0x80000008))
> - return boot_cpu_data.x86_phys_bits;

Removing this entirely will break CPUs that don't support leaf 0x80000008.
>From a VMX perspective, I'm pretty sure all Intel hardware that supports
VMX is guaranteed to support 0x80000008, but I've no idea about SVM or any
non-Intel CPU, and not supporting 0x80000008 in a virtual machine is
technically legal/possible. We conditioned doing CPUID on TME because TME
would be reported as supported iff 0x80000008 existed.

The extra bit of paranoia doesn't cost much, so play it safe? E.g.:

if (unlikely(boot_cpu_data.extended_cpuid_level < 0x80000008)) {
WARN_ON_ONCE(boot_cpu_has(X86_FEATURE_TME) || SME?);
return boot_cpu_data.x86_phys_bits;
}

return cpuid_eax(0x80000008) & 0xff;

> -
> return cpuid_eax(0x80000008) & 0xff;
> }
>
> --
> 1.8.3.1
>