Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Subject: Re: [PATCH v15 07/14] KVM: VMX: Emulate reads and writes to CET MSRs
To:     Yang Weijiang <weijiang.yang@intel.com>, seanjc@google.com,
        jmattson@google.com, kvm@vger.kernel.org,
        linux-kernel@vger.kernel.org
Cc:     yu.c.zhang@linux.intel.com
References: <20210203113421.5759-1-weijiang.yang@intel.com>
 <20210203113421.5759-8-weijiang.yang@intel.com>
From:   Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <6a743d57-8128-b5db-ddc1-9dd4c4c1004e@redhat.com>
Date:   Wed, 3 Feb 2021 12:57:41 +0100
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101
 Thunderbird/78.6.0
MIME-Version: 1.0
In-Reply-To: <20210203113421.5759-8-weijiang.yang@intel.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 03/02/21 12:34, Yang Weijiang wrote:
> MSRs that are switched through XSAVES are especially annoying due to 
> the possibility of the kernel's FPU being used in IRQ context. Disable 
> IRQs and ensure the guest's FPU state is loaded when accessing such MSRs.

Good catch!  This should be in x86.h and named kvm_get/set_xsave_msr 
because it's not VMX specific.  The commit message should also be there 
as a comment.

In addition,

> +	case MSR_IA32_S_CET:
> +		if (!cet_is_control_msr_accessible(vcpu, msr_info))
> +			return 1;
> +		msr_info->data = vmcs_readl(GUEST_S_CET);
> +		break;
> +	case MSR_IA32_U_CET:
> +		if (!cet_is_control_msr_accessible(vcpu, msr_info))
> +			return 1;
> +		vmx_get_xsave_msr(msr_info);
> +		break;

these two might as well be the same "case" for symmetry with the 
handling of WRMSR.

I've fixed this up locally, since these patches will not be pushed to 
Linus until the corresponding bare metal support is there.

Paolo