[08/22] KVM: nVMX: Cache shadow vmcs12 on VMEntry and flush to memory on VMExit

Commit Message

Paolo Bonzini July 18, 2018, 5:57 p.m. UTC

On 18/07/2018 19:40, Paolo Bonzini wrote:
> On 23/06/2018 01:35, Liran Alon wrote:
>> @@ -12392,6 +12455,17 @@ static void nested_vmx_vmexit(struct kvm_vcpu *vcpu, u32 exit_reason,
>>  			prepare_vmcs12(vcpu, vmcs12, exit_reason, exit_intr_info,
>>  				       exit_qualification);
>>  
>> +		/*
>> +		 * Must happen outside of sync_vmcs12() as it will
>> +		 * also be used to capture vmcs12 cache as part of
>> +		 * capturing nVMX state for snapshot (migration).
>> +		 *
>> +		 * Otherwise, this flush will dirty guest memory at a
>> +		 * point it is already assumed by user-space to be
>> +		 * immutable.
>> +		 */
>> +		nested_flush_cached_shadow_vmcs12(vcpu, vmcs12);
>> +
>>  		if (nested_vmx_store_msr(vcpu, vmcs12->vm_exit_msr_store_addr,
>>  					 vmcs12->vm_exit_msr_store_count))
>>  			nested_vmx_abort(vcpu, VMX_ABORT_SAVE_GUEST_MSR_FAIL);
>>
> 
> So I assume this would this be another page at the end of the nested
> state.  Do you already have a patch for that?
> 

Should be something like this :



The only question I have, should shadow_vmcs12 be included if we're
not in guest mode?  Above I'm assuming no, because it should be flushed
to memory on vmexit (including SMM entry).

Thanks,

Paolo

Patch

diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index efe3205feb48..41ca12cbd82b 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -13152,9 +13152,15 @@  static int vmx_get_nested_state(struct kvm_vcpu *vcpu,
 		kvm_state.vmx.vmxon_pa = vmx->nested.vmxon_ptr;
 		kvm_state.vmx.vmcs_pa = vmx->nested.current_vmptr;
 
-		if (vmx->nested.current_vmptr != -1ull)
+		if (vmx->nested.current_vmptr != -1ull) {
 			kvm_state.size += VMCS12_SIZE;
 
+			if (is_guest_mode(vcpu) &&
+			    nested_cpu_has_shadow_vmcs(vmcs12) &&
+			    vmcs12->vmcs_link_pointer != -1ull)
+				kvm_state.size += VMCS12_SIZE;
+		}
+
 		if (vmx->nested.smm.vmxon)
 			kvm_state.vmx.smm.flags |= KVM_STATE_NESTED_SMM_VMXON;
 
@@ -13193,6 +13198,13 @@  static int vmx_get_nested_state(struct kvm_vcpu *vcpu,
 	if (copy_to_user(user_kvm_nested_state->data, vmcs12, sizeof(*vmcs12)))
 		return -EFAULT;
 
+	if (nested_cpu_has_shadow_vmcs(vmcs12) &&
+	    vmcs12->vmcs_link_pointer != -1ull) {
+		if (copy_to_user(user_kvm_nested_state->data + VMCS12_SIZE,
+				 get_shadow_vmcs12(vcpu), sizeof(*vmcs12)))
+			return -EFAULT;
+	}
+
 out:
 	return kvm_state.size;
 }
@@ -13244,6 +13256,23 @@  static int vmx_set_nested_state(struct kvm_vcpu *vcpu,
 	    vmcs12->hdr.shadow_vmcs)
 		return -EINVAL;
 
+	if ((kvm_state->flags & KVM_STATE_NESTED_GUEST_MODE) &&
+	    nested_cpu_has_shadow_vmcs(vmcs12) &&
+	    vmcs12->vmcs_link_pointer != -1ull) {
+		struct vmcs12 *shadow_vmcs12 = get_shadow_vmcs12(vcpu);
+		if (kvm_state->size < sizeof(kvm_state) + 2 * sizeof(*vmcs12))
+			return -EINVAL;
+
+		if (copy_from_user(shadow_vmcs12,
+				   user_kvm_nested_state->data + VMCS12_SIZE,
+				   sizeof(*vmcs12)))
+			return -EFAULT;
+
+		if (shadow_vmcs12->hdr.revision_id != VMCS12_REVISION ||
+		    !shadow_vmcs12->hdr.shadow_vmcs)
+			return -EINVAL;
+	}
+
 	vmx_leave_nested(vcpu);
 	if (kvm_state->vmx.vmxon_pa == -1ull)
 		return 0;