Re: [PATCH v10 07/28] KVM: VMX: Initialize VMCS FRED fields

[Binbin Wu]

On 9/12/2026 5:36 AM, Sohil Mehta wrote:
> From: "Xin Li (Intel)" <xin@xxxxxxxxx>
> 
> Initialize host VMCS FRED fields with host FRED MSRs' value and
> guest VMCS FRED fields to 0.
> 
> FRED CPU state is managed in 9 new FRED MSRs:
>         IA32_FRED_CONFIG,
>         IA32_FRED_STKLVLS,
>         IA32_FRED_RSP0,
>         IA32_FRED_RSP1,
>         IA32_FRED_RSP2,
>         IA32_FRED_RSP3,
>         IA32_FRED_SSP1,
>         IA32_FRED_SSP2,
>         IA32_FRED_SSP3,
> as well as a few existing CPU registers and MSRs:
>         CR4.FRED,
>         IA32_STAR,
>         IA32_KERNEL_GS_BASE,
>         IA32_PL0_SSP (also known as IA32_FRED_SSP0).
> 
> CR4, IA32_KERNEL_GS_BASE and IA32_STAR are already well managed.
> Except IA32_FRED_RSP0 and IA32_FRED_SSP0, all other FRED CPU state
> MSRs have corresponding VMCS fields in both the host-state and
> guest-state areas.  So KVM just needs to initialize them, and with
> proper VM entry/exit FRED controls, a FRED CPU will keep tracking
> host and guest FRED CPU state in VMCS automatically.
> 
> Signed-off-by: Xin Li (Intel) <xin@xxxxxxxxx>
> Signed-off-by: Sohil Mehta <sohil.mehta@xxxxxxxxx>
> Reviewed-by: Chao Gao <chao.gao@xxxxxxxxx>
> Reviewed-by: Binbin Wu <binbin.wu@xxxxxxxxxxxxxxx>
> ---
> v10:
>  - s/rdmsrl/rdmsrq (Chao)
>  - Remove a terse comment which doesn't help at all (Sean).
>  - Replace direct, raw use of __this_cpu_ist_top_va() with a self-
>    explanatory this_cpu_fred_rsp() helper for better readability (Sean).
>  - Add cpu_feature_enabled(X86_FEATURE_FRED) to patch out accesses to
>    VMCS FRED fields (Sean).
>  - Remove "#ifdef CONFIG_X86_64" after switching to this_cpu_fred_rsp()
>    (Sean).

[...]

> diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
> index 2a2218e0983b..dc80ca5f804d 100644
> --- a/arch/x86/kvm/vmx/vmx.c
> +++ b/arch/x86/kvm/vmx/vmx.c
> @@ -1562,6 +1562,12 @@ void vmx_vcpu_load_vmcs(struct kvm_vcpu *vcpu, int cpu)
>  				    (unsigned long)(cpu_entry_stack(cpu) + 1));
>  		}
>  
> +		if (cpu_feature_enabled(X86_FEATURE_FRED) && kvm_cpu_cap_has(X86_FEATURE_FRED)) {

Nit:
I was wondering why cpu_feature_enabled(X86_FEATURE_FRED) was needed for a while.
Sorry I missed the version history :(

This seems to be the first time this cpu_feature_enabled() && kvm_cpu_cap_has() combination is used,
maybe have a mention in the changelog?

> +			vmcs_write64(HOST_IA32_FRED_RSP1, this_cpu_fred_rsp(FRED_STACK_LEVEL_1));
> +			vmcs_write64(HOST_IA32_FRED_RSP2, this_cpu_fred_rsp(FRED_STACK_LEVEL_2));
> +			vmcs_write64(HOST_IA32_FRED_RSP3, this_cpu_fred_rsp(FRED_STACK_LEVEL_3));
> +		}
> +
>  		vmx->loaded_vmcs->cpu = cpu;
>  	}
>  }
> @@ -4500,6 +4506,17 @@ void vmx_set_constant_host_state(struct vcpu_vmx *vmx)
>  	 */
>  	vmcs_write16(HOST_DS_SELECTOR, 0);
>  	vmcs_write16(HOST_ES_SELECTOR, 0);
> +
> +	if (cpu_feature_enabled(X86_FEATURE_FRED) && kvm_cpu_cap_has(X86_FEATURE_FRED)) {
> +		/* FRED CONFIG and STKLVLS are the same on all CPUs */
> +		vmcs_write64(HOST_IA32_FRED_CONFIG, kvm_host.fred_config);
> +		vmcs_write64(HOST_IA32_FRED_STKLVLS, kvm_host.fred_stklvls);
> +
> +		/* Linux doesn't support kernel shadow stacks, thus SSPs are 0s */
> +		vmcs_write64(HOST_IA32_FRED_SSP1, 0);
> +		vmcs_write64(HOST_IA32_FRED_SSP2, 0);
> +		vmcs_write64(HOST_IA32_FRED_SSP3, 0);
> +	}
>  #else
>  	vmcs_write16(HOST_DS_SELECTOR, __KERNEL_DS);  /* 22.2.4 */
>  	vmcs_write16(HOST_ES_SELECTOR, __KERNEL_DS);  /* 22.2.4 */
> @@ -5030,6 +5047,17 @@ static void init_vmcs(struct vcpu_vmx *vmx)
>  	}
>  
>  	vmx_setup_uret_msrs(vmx);
> +
> +	if (cpu_feature_enabled(X86_FEATURE_FRED) && kvm_cpu_cap_has(X86_FEATURE_FRED)) {
> +		vmcs_write64(GUEST_IA32_FRED_CONFIG, 0);
> +		vmcs_write64(GUEST_IA32_FRED_RSP1, 0);
> +		vmcs_write64(GUEST_IA32_FRED_RSP2, 0);
> +		vmcs_write64(GUEST_IA32_FRED_RSP3, 0);
> +		vmcs_write64(GUEST_IA32_FRED_STKLVLS, 0);
> +		vmcs_write64(GUEST_IA32_FRED_SSP1, 0);
> +		vmcs_write64(GUEST_IA32_FRED_SSP2, 0);
> +		vmcs_write64(GUEST_IA32_FRED_SSP3, 0);
> +	}
>  }
>  
>  static void __vmx_vcpu_reset(struct kvm_vcpu *vcpu)
> @@ -8883,6 +8911,11 @@ __init int vmx_hardware_setup(void)
>  
>  	kvm_caps.inapplicable_quirks &= ~KVM_X86_QUIRK_IGNORE_GUEST_PAT;
>  
> +	if (cpu_feature_enabled(X86_FEATURE_FRED) && kvm_cpu_cap_has(X86_FEATURE_FRED)) {
> +		rdmsrq(MSR_IA32_FRED_CONFIG, kvm_host.fred_config);
> +		rdmsrq(MSR_IA32_FRED_STKLVLS, kvm_host.fred_stklvls);
> +	}
> +
>  	return 0;
>  }
>