2013-04-15 12:46:45

by Paolo Bonzini

Subject: [PATCH] kvm: nVMX: check vmcs12 for valid activity state

KVM does not use the activity state VMCS field, and does not support
it in nested VMX either (the corresponding bits in the misc VMX feature
MSR are zero). Fail entry if the activity state is set to anything but
"active".

Signed-off-by: Paolo Bonzini <[email protected]>
---
arch/x86/kvm/vmx.c | 5 +++++
1 file changed, 5 insertions(+)

diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index c75c25d..589e79a 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -7223,6 +7223,11 @@ static int nested_vmx_run(struct kvm_vcpu *vcpu, bool launch)
return 1;
}

+ if (vmcs12->guest_activity_state != 0) {
+ nested_vmx_failValid(vcpu, VMXERR_ENTRY_INVALID_CONTROL_FIELD);
+ return 1;
+ }
+
if ((vmcs12->cpu_based_vm_exec_control & CPU_BASED_USE_MSR_BITMAPS) &&
!IS_ALIGNED(vmcs12->msr_bitmap, PAGE_SIZE)) {
/*TODO: Also verify bits beyond physical address width are 0*/
--
1.8.1.4

2013-04-15 12:54:18

by Jan Kiszka

Subject: Re: [PATCH] kvm: nVMX: check vmcs12 for valid activity state

On 2013-04-15 14:46, Paolo Bonzini wrote:
> KVM does not use the activity state VMCS field, and does not support
> it in nested VMX either (the corresponding bits in the misc VMX feature
> MSR are zero). Fail entry if the activity state is set to anything but
> "active".

Ah, that indeed simplifies the issue.

>
> Signed-off-by: Paolo Bonzini <[email protected]>
> ---
> arch/x86/kvm/vmx.c | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
> index c75c25d..589e79a 100644
> --- a/arch/x86/kvm/vmx.c
> +++ b/arch/x86/kvm/vmx.c
> @@ -7223,6 +7223,11 @@ static int nested_vmx_run(struct kvm_vcpu *vcpu, bool launch)
> return 1;
> }
>
> + if (vmcs12->guest_activity_state != 0) {

GUEST_ACTIVITY_ACTIVE, please.

> + nested_vmx_failValid(vcpu, VMXERR_ENTRY_INVALID_CONTROL_FIELD);
> + return 1;
> + }
> +
> if ((vmcs12->cpu_based_vm_exec_control & CPU_BASED_USE_MSR_BITMAPS) &&
> !IS_ALIGNED(vmcs12->msr_bitmap, PAGE_SIZE)) {
> /*TODO: Also verify bits beyond physical address width are 0*/
>

That means we also do not need to read and write the corresponding VMCS
field on L1/L2 transitions - avoids any risk of setting it to an
unsupported value. So I would fold it into this change.

Jan

--
Siemens AG, Corporate Technology, CT RTC ITP SDP-DE
Corporate Competence Center Embedded Linux