Re: [PATCH v12 10/10] KVM: x86: Enable CET virtualization and advertise CET to userspace

[Sean Christopherson]

On Wed, May 06, 2020 at 04:21:09PM +0800, Yang Weijiang wrote:
> Set the feature bits so that CET capabilities can be seen in guest via
> CPUID enumeration. Add CR4.CET bit support in order to allow guest set CET
> master control bit(CR4.CET).
> 
> Signed-off-by: Yang Weijiang <weijiang.yang@xxxxxxxxx>
> ---
>  arch/x86/include/asm/kvm_host.h | 3 ++-
>  arch/x86/kvm/cpuid.c            | 5 +++--
>  2 files changed, 5 insertions(+), 3 deletions(-)
> 
> diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
> index f68c825e94ad..21f3c89d8c70 100644
> --- a/arch/x86/include/asm/kvm_host.h
> +++ b/arch/x86/include/asm/kvm_host.h
> @@ -95,7 +95,8 @@
>  			  | X86_CR4_PGE | X86_CR4_PCE | X86_CR4_OSFXSR | X86_CR4_PCIDE \
>  			  | X86_CR4_OSXSAVE | X86_CR4_SMEP | X86_CR4_FSGSBASE \
>  			  | X86_CR4_OSXMMEXCPT | X86_CR4_LA57 | X86_CR4_VMXE \
> -			  | X86_CR4_SMAP | X86_CR4_PKE | X86_CR4_UMIP))
> +			  | X86_CR4_SMAP | X86_CR4_PKE | X86_CR4_UMIP \
> +			  | X86_CR4_CET))
>  
>  #define CR8_RESERVED_BITS (~(unsigned long)X86_CR8_TPR)
>  
> diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c
> index 984ab2b395b3..333a9e0d7cdf 100644
> --- a/arch/x86/kvm/cpuid.c
> +++ b/arch/x86/kvm/cpuid.c
> @@ -344,7 +344,8 @@ void kvm_set_cpu_caps(void)
>  		F(AVX512VBMI) | F(LA57) | 0 /*PKU*/ | 0 /*OSPKE*/ | F(RDPID) |
>  		F(AVX512_VPOPCNTDQ) | F(UMIP) | F(AVX512_VBMI2) | F(GFNI) |
>  		F(VAES) | F(VPCLMULQDQ) | F(AVX512_VNNI) | F(AVX512_BITALG) |
> -		F(CLDEMOTE) | F(MOVDIRI) | F(MOVDIR64B) | 0 /*WAITPKG*/
> +		F(CLDEMOTE) | F(MOVDIRI) | F(MOVDIR64B) | 0 /*WAITPKG*/ |
> +		F(SHSTK)
>  	);
>  	/* Set LA57 based on hardware capability. */
>  	if (cpuid_ecx(7) & F(LA57))
> @@ -353,7 +354,7 @@ void kvm_set_cpu_caps(void)
>  	kvm_cpu_cap_mask(CPUID_7_EDX,
>  		F(AVX512_4VNNIW) | F(AVX512_4FMAPS) | F(SPEC_CTRL) |
>  		F(SPEC_CTRL_SSBD) | F(ARCH_CAPABILITIES) | F(INTEL_STIBP) |
> -		F(MD_CLEAR) | F(AVX512_VP2INTERSECT) | F(FSRM)
> +		F(MD_CLEAR) | F(AVX512_VP2INTERSECT) | F(FSRM) | F(IBT)

SHSTK and IBT need to be disabled in vmx_set_cpu_caps() if unrestricted guest
is disabled.  CET won't play nice with emulating arbitrary instructions, e.g.
KVM doesn't enforce ENDBR and doesn't keep SSP up-to-date (and no one is
advocating fully emulating CET).

Paolo also floated the idea of providing a reduced opcode set, e.g. only I/O,
MOV, and ALU instructions, but I don't think that needs to be done in the
initial CET enabling as it's more of a defense-in-depth than a functional
requirement.

No need to respin a new series just for this, it can wait until I've looked
through this version.

Original thread: https://lkml.kernel.org/r/20200515161919.29249-1-pbonzini@xxxxxxxxxx

>  	);
>  
>  	/* TSC_ADJUST and ARCH_CAPABILITIES are emulated in software. */
> -- 
> 2.17.2
>