Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <1517856451-2932-1-git-send-email-karahmed@amazon.de> <1517856451-2932-6-git-send-email-karahmed@amazon.de>
In-Reply-To: <1517856451-2932-6-git-send-email-karahmed@amazon.de>
From:   Jim Mattson <jmattson@google.com>
Date:   Mon, 05 Feb 2018 22:15:01 +0000
Message-ID: <CALMp9eScXeCi6J4kuktdZmmOg2oDyyatpzOdVmUSso7F8FU+zA@mail.gmail.com>
Subject: Re: [RFC 05/12] KVM/VMX: Use the new host mapping API for mapping
 nested vmptr
To:     KarimAllah Ahmed <karahmed@amazon.de>
Cc:     LKML <linux-kernel@vger.kernel.org>,
        kvm list <kvm@vger.kernel.org>,
        Paolo Bonzini <pbonzini@redhat.com>,
        =?UTF-8?B?UmFkaW0gS3LEjW3DocWZ?= <rkrcmar@redhat.com>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Mon, Feb 5, 2018 at 10:49 AM KarimAllah Ahmed <karahmed@amazon.de> wrote:


> @@ -7410,19 +7410,17 @@ static int handle_vmon(struct kvm_vcpu *vcpu)
>                  return kvm_skip_emulated_instruction(vcpu);
>          }

> -       page = kvm_vcpu_gpa_to_page(vcpu, vmptr);
> -       if (is_error_page(page)) {
> +       if (!kvm_vcpu_gpa_to_host_mapping(vcpu, vmptr, &mapping, true)) {
>                  nested_vmx_failInvalid(vcpu);
>                  return kvm_skip_emulated_instruction(vcpu);
>          }
> -       if (*(u32 *)kmap(page) != VMCS12_REVISION) {
> -               kunmap(page);
> -               kvm_release_page_clean(page);
> +       if (*(u32 *)mapping.kaddr != VMCS12_REVISION) {
> +               kvm_release_host_mapping(&mapping, false);
>                  nested_vmx_failInvalid(vcpu);
>                  return kvm_skip_emulated_instruction(vcpu);
>          }
> -       kunmap(page);
> -       kvm_release_page_clean(page);
> +
> +       kvm_release_host_mapping(&mapping, false);

Why go through this explicit mapping/release dance? Why not just:

uint32_t revision;
...
if (kvm_read_guest(vcpu->kvm, vmptr, &revision, sizeof(revision)) ||
     revision != VMCS12_REVISION) {
         nested_vmx_failInvalid(vcpu);
         return kvm_skip_emulated_instruction(vcpu);
}