linux-kernel - Re: [PATCH v5 6/7] KVM: VMX: Expose PKS to guest

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <YYmXO2WQpydWVro0@google.com>
Date:   Mon, 8 Nov 2021 21:31:39 +0000
From:   Sean Christopherson <seanjc@...gle.com>
To:     Chenyi Qiang <chenyi.qiang@...el.com>
Cc:     Paolo Bonzini <pbonzini@...hat.com>,
        Vitaly Kuznetsov <vkuznets@...hat.com>,
        Wanpeng Li <wanpengli@...cent.com>,
        Jim Mattson <jmattson@...gle.com>,
        Joerg Roedel <joro@...tes.org>,
        Xiaoyao Li <xiaoyao.li@...el.com>, kvm@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v5 6/7] KVM: VMX: Expose PKS to guest

On Wed, Aug 11, 2021, Chenyi Qiang wrote:
> diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c
> index 739be5da3bca..dbee0d639db3 100644
> --- a/arch/x86/kvm/cpuid.c
> +++ b/arch/x86/kvm/cpuid.c
> @@ -458,7 +458,7 @@ void kvm_set_cpu_caps(void)
>  		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(SGX_LC) | F(BUS_LOCK_DETECT)
> +		F(SGX_LC) | F(BUS_LOCK_DETECT) | 0 /*PKS*/

...

>  	);
>  	/* Set LA57 based on hardware capability. */
>  	if (cpuid_ecx(7) & F(LA57))

...

> @@ -7311,6 +7312,14 @@ static __init void vmx_set_cpu_caps(void)
>  
>  	if (cpu_has_vmx_waitpkg())
>  		kvm_cpu_cap_check_and_set(X86_FEATURE_WAITPKG);
> +
> +	/*
> +	 * PKS is not yet implemented for shadow paging.
> +	 * If not support VM_{ENTRY, EXIT}_LOAD_IA32_PKRS,
> +	 * don't expose the PKS as well.
> +	 */
> +	if (enable_ept && cpu_has_load_ia32_pkrs())
> +		kvm_cpu_cap_check_and_set(X86_FEATURE_PKS);

I would rather handle the !TDP case in cpuid.c alongside the PKU.  The decision
to not support Protection Keys with legacy shadow paging is an x86 decision, not
a VMX decision.

And VMX's extra restriction on the VMCS support should not bleed into common x86.

Can you also opportunistically update the comment (see below) to explain _why_
OSPKE needs to be enabled in order to advertise PKU?

Thanks!

diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c
index 2d70edb0f323..c4ed6881857c 100644
--- a/arch/x86/kvm/cpuid.c
+++ b/arch/x86/kvm/cpuid.c
@@ -439,18 +439,23 @@ void kvm_set_cpu_caps(void)
                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(SGX_LC) | F(BUS_LOCK_DETECT)
+               F(SGX_LC) | F(BUS_LOCK_DETECT) | F(PKS)
        );
        /* Set LA57 based on hardware capability. */
        if (cpuid_ecx(7) & F(LA57))
                kvm_cpu_cap_set(X86_FEATURE_LA57);

        /*
-        * PKU not yet implemented for shadow paging and requires OSPKE
-        * to be set on the host. Clear it if that is not the case
+        * Protection Keys are not supported for shadow paging.  PKU further
+        * requires OSPKE to be set on the host in order to use {RD,WR}PKRU to
+        * save/restore the guests PKRU.
         */
-       if (!tdp_enabled || !boot_cpu_has(X86_FEATURE_OSPKE))
+       if (!tdp_enabled) {
                kvm_cpu_cap_clear(X86_FEATURE_PKU);
+               kvm_cpu_cap_clear(X86_FEATURE_PKS);
+       } else if (!boot_cpu_has(X86_FEATURE_OSPKE)) {
+               kvm_cpu_cap_clear(X86_FEATURE_PKU);
+       }

        kvm_cpu_cap_mask(CPUID_7_EDX,
                F(AVX512_4VNNIW) | F(AVX512_4FMAPS) | F(SPEC_CTRL) |


and then vmx.c only needs to handle clearing PKS when the VMCS controls aren't
available.