MIME-Version: 1.0
In-Reply-To: <1515434925-10250-5-git-send-email-pbonzini@redhat.com>
References: <1515434925-10250-1-git-send-email-pbonzini@redhat.com> <1515434925-10250-5-git-send-email-pbonzini@redhat.com>
From: Jim Mattson <jmattson@google.com>
Date: Mon, 8 Jan 2018 11:36:51 -0800
Message-ID: <CALMp9eThEzb5UR_vwFNrUmQ+--7P39r65c-t-sAS7TKPPw87Jw@mail.gmail.com>
Subject: Re: [PATCH 4/7] kvm: vmx: Set IBPB when running a different VCPU
To: Paolo Bonzini <pbonzini@redhat.com>
Cc: LKML <linux-kernel@vger.kernel.org>,
        kvm list <kvm@vger.kernel.org>, aliguori@amazon.com,
        Tom Lendacky <thomas.lendacky@amd.com>, dwmw@amazon.co.uk,
        bp@alien8.de, Tim Chen <tim.c.chen@linux.intel.com>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org

Shouldn't there be an IBPB on *any* context switch away from a VCPU
thread, even if it is to a non-VCPU thread?

On Mon, Jan 8, 2018 at 10:08 AM, Paolo Bonzini <pbonzini@redhat.com> wrote:
> From: Tim Chen <tim.c.chen@linux.intel.com>
>
> Ensure an IBPB (Indirect branch prediction barrier) before every VCPU
> switch.
>
> Signed-off-by: Tim Chen <tim.c.chen@linux.intel.com>
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> ---
>  arch/x86/kvm/vmx.c | 9 +++++++++
>  1 file changed, 9 insertions(+)
>
> diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
> index d00bcad7336e..bf127c570675 100644
> --- a/arch/x86/kvm/vmx.c
> +++ b/arch/x86/kvm/vmx.c
> @@ -2375,6 +2375,8 @@ static void vmx_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
>         if (per_cpu(current_vmcs, cpu) != vmx->loaded_vmcs->vmcs) {
>                 per_cpu(current_vmcs, cpu) = vmx->loaded_vmcs->vmcs;
>                 vmcs_load(vmx->loaded_vmcs->vmcs);
> +               if (have_spec_ctrl)
> +                       wrmsrl(MSR_IA32_PRED_CMD, FEATURE_SET_IBPB);
>         }
>
>         if (!already_loaded) {
> @@ -4029,6 +4031,13 @@ static void free_loaded_vmcs(struct loaded_vmcs *loaded_vmcs)
>         free_vmcs(loaded_vmcs->vmcs);
>         loaded_vmcs->vmcs = NULL;
>         WARN_ON(loaded_vmcs->shadow_vmcs != NULL);
> +
> +       /*
> +        * The VMCS could be recycled, causing a false negative in
> +        * vmx_vcpu_load; block speculative execution.
> +        */
> +       if (have_spec_ctrl)
> +               wrmsrl(MSR_IA32_PRED_CMD, FEATURE_SET_IBPB);
>  }
>
>  static void vmx_nested_free_vmcs02(struct vcpu_vmx *vmx)
> --
> 1.8.3.1
>
>