linux-kernel - Re: [PATCH RFC 2/3] x86/kvm/hyper-v: move VMX controls sanitization out of nested_enable

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: <6c4bdb57-08fb-2c2d-9234-b7efffeb72ed@redhat.com>
Date:   Sun, 19 Jan 2020 09:54:44 +0100
From:   Paolo Bonzini <pbonzini@...hat.com>
To:     Vitaly Kuznetsov <vkuznets@...hat.com>, kvm@...r.kernel.org
Cc:     Sean Christopherson <sean.j.christopherson@...el.com>,
        Jim Mattson <jmattson@...gle.com>,
        linux-kernel@...r.kernel.org, Liran Alon <liran.alon@...cle.com>,
        Roman Kagan <rkagan@...tuozzo.com>
Subject: Re: [PATCH RFC 2/3] x86/kvm/hyper-v: move VMX controls sanitization
 out of nested_enable_evmcs()

On 15/01/20 18:10, Vitaly Kuznetsov wrote:
> With fine grained VMX feature enablement QEMU>=4.2 tries to do KVM_SET_MSRS
> with default (matching CPU model) values and in case eVMCS is also enabled,
> fails.
> 
> It would be possible to drop VMX feature filtering completely and make
> this a guest's responsibility: if it decides to use eVMCS it should know
> which fields are available and which are not. Hyper-V mostly complies to
> this, however, there is at least one problematic control:
> SECONDARY_EXEC_VIRTUALIZE_APIC_ACCESSES
> which Hyper-V enables. As there is no 'apic_addr_field' in eVMCS, we
> fail to handle this properly in KVM. It is unclear how this is supposed
> to work, genuine Hyper-V doesn't expose the control so it is possible that
> this is just a bug (in Hyper-V).

Yes, it most likely is and it would be nice if Microsoft fixed it, but I
guess we're stuck with it for existing Windows versions.  Well, for one
we found a bug in Hyper-V and not the converse. :)

There is a problem with this approach, in that we're stuck with it
forever due to live migration.  But I guess if in the future eVMCS v2
adds an apic_address field we can limit the hack to eVMCS v1.  Another
possibility is to use the quirks mechanism but it's overkill for now.

Unless there are objections, I plan to apply these patches.

Paolo


> Move VMX controls sanitization from nested_enable_evmcs() to vmx_get_msr(),
> this allows userspace to keep setting controls it wants and at the same
> time hides them from the guest.
> 
> Signed-off-by: Vitaly Kuznetsov <vkuznets@...hat.com>
> ---
>  arch/x86/kvm/vmx/evmcs.c | 38 ++++++++++++++++++++++++++++++++------
>  arch/x86/kvm/vmx/evmcs.h |  1 +
>  arch/x86/kvm/vmx/vmx.c   | 10 ++++++++--
>  3 files changed, 41 insertions(+), 8 deletions(-)
> 
> diff --git a/arch/x86/kvm/vmx/evmcs.c b/arch/x86/kvm/vmx/evmcs.c
> index 89c3e0caf39f..b5d6582ba589 100644
> --- a/arch/x86/kvm/vmx/evmcs.c
> +++ b/arch/x86/kvm/vmx/evmcs.c
> @@ -346,6 +346,38 @@ uint16_t nested_get_evmcs_version(struct kvm_vcpu *vcpu)
>         return 0;
>  }
>  
> +void nested_evmcs_filter_control_msr(u32 msr_index, u64 *pdata)
> +{
> +	u32 ctl_low = (u32)*pdata, ctl_high = (u32)(*pdata >> 32);
> +	/*
> +	 * Enlightened VMCS doesn't have certain fields, make sure we don't
> +	 * expose unsupported controls to L1.
> +	 */
> +
> +	switch (msr_index) {
> +	case MSR_IA32_VMX_PINBASED_CTLS:
> +	case MSR_IA32_VMX_TRUE_PINBASED_CTLS:
> +		ctl_high &= ~EVMCS1_UNSUPPORTED_PINCTRL;
> +		break;
> +	case MSR_IA32_VMX_EXIT_CTLS:
> +	case MSR_IA32_VMX_TRUE_EXIT_CTLS:
> +		ctl_high &= ~EVMCS1_UNSUPPORTED_VMEXIT_CTRL;
> +		break;
> +	case MSR_IA32_VMX_ENTRY_CTLS:
> +	case MSR_IA32_VMX_TRUE_ENTRY_CTLS:
> +		ctl_high &= ~EVMCS1_UNSUPPORTED_VMENTRY_CTRL;
> +		break;
> +	case MSR_IA32_VMX_PROCBASED_CTLS2:
> +		ctl_high &= ~EVMCS1_UNSUPPORTED_2NDEXEC;
> +		break;
> +	case MSR_IA32_VMX_VMFUNC:
> +		ctl_low &= ~EVMCS1_UNSUPPORTED_VMFUNC;
> +		break;
> +	}
> +
> +	*pdata = ctl_low | ((u64)ctl_high << 32);
> +}
> +
>  int nested_enable_evmcs(struct kvm_vcpu *vcpu,
>  			uint16_t *vmcs_version)
>  {
> @@ -356,11 +388,5 @@ int nested_enable_evmcs(struct kvm_vcpu *vcpu,
>  	if (vmcs_version)
>  		*vmcs_version = nested_get_evmcs_version(vcpu);
>  
> -	vmx->nested.msrs.pinbased_ctls_high &= ~EVMCS1_UNSUPPORTED_PINCTRL;
> -	vmx->nested.msrs.entry_ctls_high &= ~EVMCS1_UNSUPPORTED_VMENTRY_CTRL;
> -	vmx->nested.msrs.exit_ctls_high &= ~EVMCS1_UNSUPPORTED_VMEXIT_CTRL;
> -	vmx->nested.msrs.secondary_ctls_high &= ~EVMCS1_UNSUPPORTED_2NDEXEC;
> -	vmx->nested.msrs.vmfunc_controls &= ~EVMCS1_UNSUPPORTED_VMFUNC;
> -
>  	return 0;
>  }
> diff --git a/arch/x86/kvm/vmx/evmcs.h b/arch/x86/kvm/vmx/evmcs.h
> index 07ebf6882a45..b88d9807a796 100644
> --- a/arch/x86/kvm/vmx/evmcs.h
> +++ b/arch/x86/kvm/vmx/evmcs.h
> @@ -201,5 +201,6 @@ bool nested_enlightened_vmentry(struct kvm_vcpu *vcpu, u64 *evmcs_gpa);
>  uint16_t nested_get_evmcs_version(struct kvm_vcpu *vcpu);
>  int nested_enable_evmcs(struct kvm_vcpu *vcpu,
>  			uint16_t *vmcs_version);
> +void nested_evmcs_filter_control_msr(u32 msr_index, u64 *pdata);
>  
>  #endif /* __KVM_X86_VMX_EVMCS_H */
> diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
> index e3394c839dea..8eb74618b8d8 100644
> --- a/arch/x86/kvm/vmx/vmx.c
> +++ b/arch/x86/kvm/vmx/vmx.c
> @@ -1849,8 +1849,14 @@ static int vmx_get_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
>  	case MSR_IA32_VMX_BASIC ... MSR_IA32_VMX_VMFUNC:
>  		if (!nested_vmx_allowed(vcpu))
>  			return 1;
> -		return vmx_get_vmx_msr(&vmx->nested.msrs, msr_info->index,
> -				       &msr_info->data);
> +		if (vmx_get_vmx_msr(&vmx->nested.msrs, msr_info->index,
> +				    &msr_info->data))
> +			return 1;
> +		if (!msr_info->host_initiated &&
> +		    vmx->nested.enlightened_vmcs_enabled)
> +			nested_evmcs_filter_control_msr(msr_info->index,
> +							&msr_info->data);
> +		break;
>  	case MSR_IA32_RTIT_CTL:
>  		if (pt_mode != PT_MODE_HOST_GUEST)
>  			return 1;
>