Skip to contentSkip to conversation

KVM: nVMX: fix comments of handle_vmon()

PATCH Applied Yu Zhang 3 participants 6 messages Sep 8, 2021 – Sep 21, 2021

Subsystems (3)
lore
Share & tools
lore

Yu Zhang

author
#0 lore raw

PATCH 1/1 KVM: nVMX: fix comments of handle_vmon()

"VMXON pointer" is saved in vmx->nested.vmxon_ptr since commit 3573e22cfeca ("KVM: nVMX: additional checks on vmxon region"). Also, handle_vmptrld() & handle_vmclear() now have logic to check the VMCS pointer against the VMXON pointer.

So just remove the obsolete comments of handle_vmon().

---
 arch/x86/kvm/vmx/nested.c | 9 +--------
 1 file changed, 1 insertion(+), 8 deletions(-)
arch/x86/kvm/vmx/nested.c +1 -8
index bc6327950657..90f34f12f883 100644
@@ -4862,14 +4862,7 @@ static int enter_vmx_operation(struct kvm_vcpu *vcpu)
 	return -ENOMEM;
 }
 
-/*
- * Emulate the VMXON instruction.
- * Currently, we just remember that VMX is active, and do not save or even
- * inspect the argument to VMXON (the so-called "VMXON pointer") because we
- * do not currently need to store anything in that guest-allocated memory
- * region. Consequently, VMCLEAR and VMPTRLD also do not verify that the their
- * argument is different from the VMXON pointer (which the spec says they do).
- */
+/* Emulate the VMXON instruction. */
 static int handle_vmon(struct kvm_vcpu *vcpu)
 {
 	int ret;
Signature · 1 line
2.17.1
Reviewed-by on v1 1/1
#1 lore raw
Quoted 23 lines · Yu Zhang

"VMXON pointer" is saved in vmx->nested.vmxon_ptr since commit 3573e22cfeca ("KVM: nVMX: additional checks on vmxon region"). Also, handle_vmptrld() & handle_vmclear() now have logic to check the VMCS pointer against the VMXON pointer.

So just remove the obsolete comments of handle_vmon().

---
 arch/x86/kvm/vmx/nested.c | 9 +--------
 1 file changed, 1 insertion(+), 8 deletions(-)
arch/x86/kvm/vmx/nested.c +0 -2
index bc6327950657..90f34f12f883 100644
@@ -4862,14 +4862,7 @@ static int enter_vmx_operation(struct kvm_vcpu *vcpu)
 	return -ENOMEM;
 }
 
-/*
- * Emulate the VMXON instruction.

arch/x86/kvm/vmx/nested.c static int enter_vmx_operation(struct kvm_vcpu *vcpu) · line 4868

Quoted 10 lines
arch/x86/kvm/vmx/nested.c +1 -6
- * Currently, we just remember that VMX is active, and do not save or even
- * inspect the argument to VMXON (the so-called "VMXON pointer") because we
- * do not currently need to store anything in that guest-allocated memory
- * region. Consequently, VMCLEAR and VMPTRLD also do not verify that the their
- * argument is different from the VMXON pointer (which the spec says they do).
- */
+/* Emulate the VMXON instruction. */
 static int handle_vmon(struct kvm_vcpu *vcpu)
 {
 	int ret;

Indeed,

On a slightly related note: we don't seem to reset 'vmx->nested.vmxon_ptr' upon VMXOFF emulation; this is not a problem per se as we never access it when !vmx->nested.vmxon but I'd still suggest we do something like

arch/x86/kvm/vmx/nested.c +1 -0
index bc6327950657..8beb41d02d21 100644
@@ -290,6 +290,7 @@ static void free_nested(struct kvm_vcpu *vcpu)
 
        vmx->nested.vmxon = false;
        vmx->nested.smm.vmxon = false;
+       vmx->nested.vmxon_ptr = -1ull;
        free_vpid(vmx->nested.vpid02);
        vmx->nested.posted_intr_nv = -1;
        vmx->nested.current_vmptr = -1ull;
arch/x86/kvm/vmx/vmx.c +1 -0
index d7c5257eb5c0..2214e6bd4713 100644
@@ -6884,6 +6884,7 @@ static int vmx_create_vcpu(struct kvm_vcpu *vcpu)
 
        vcpu_setup_sgx_lepubkeyhash(vcpu);
 
+       vmx->nested.vmxon_ptr = -1ull;
        vmx->nested.posted_intr_nv = -1;
        vmx->nested.current_vmptr = -1ull;
        vmx->nested.hv_evmcs_vmptr = EVMPTR_INVALID;

to avoid issues in the future.

Signature · 1 line
Vitaly
author
#2 lore raw
Quoted 34 lines · Vitaly Kuznetsov

Indeed,

On a slightly related note: we don't seem to reset 'vmx->nested.vmxon_ptr' upon VMXOFF emulation; this is not a problem per se as we never access it when !vmx->nested.vmxon but I'd still suggest we do something like

arch/x86/kvm/vmx/nested.c +1 -0
index bc6327950657..8beb41d02d21 100644
@@ -290,6 +290,7 @@ static void free_nested(struct kvm_vcpu *vcpu)
 
        vmx->nested.vmxon = false;
        vmx->nested.smm.vmxon = false;
+       vmx->nested.vmxon_ptr = -1ull;
        free_vpid(vmx->nested.vpid02);
        vmx->nested.posted_intr_nv = -1;
        vmx->nested.current_vmptr = -1ull;
arch/x86/kvm/vmx/vmx.c +1 -0
index d7c5257eb5c0..2214e6bd4713 100644
@@ -6884,6 +6884,7 @@ static int vmx_create_vcpu(struct kvm_vcpu *vcpu)
 
        vcpu_setup_sgx_lepubkeyhash(vcpu);
 
+       vmx->nested.vmxon_ptr = -1ull;
        vmx->nested.posted_intr_nv = -1;
        vmx->nested.current_vmptr = -1ull;
        vmx->nested.hv_evmcs_vmptr = EVMPTR_INVALID;
Quoted 2 lines

to avoid issues in the future.

Thanks, Vitaly. And above suggestion sounds reasonable to me. :)

I can add this in V2 later, if no one objects.

B.R.
Yu

#3 lore raw
Quoted 35 lines · Yu Zhang
Quoted 34 lines · Vitaly Kuznetsov

Indeed,

On a slightly related note: we don't seem to reset 'vmx->nested.vmxon_ptr' upon VMXOFF emulation; this is not a problem per se as we never access it when !vmx->nested.vmxon but I'd still suggest we do something like

arch/x86/kvm/vmx/nested.c +1 -0
index bc6327950657..8beb41d02d21 100644
@@ -290,6 +290,7 @@ static void free_nested(struct kvm_vcpu *vcpu)
 
        vmx->nested.vmxon = false;
        vmx->nested.smm.vmxon = false;
+       vmx->nested.vmxon_ptr = -1ull;
        free_vpid(vmx->nested.vpid02);
        vmx->nested.posted_intr_nv = -1;
        vmx->nested.current_vmptr = -1ull;
arch/x86/kvm/vmx/vmx.c +1 -0
index d7c5257eb5c0..2214e6bd4713 100644
@@ -6884,6 +6884,7 @@ static int vmx_create_vcpu(struct kvm_vcpu *vcpu)
 
        vcpu_setup_sgx_lepubkeyhash(vcpu);
 
+       vmx->nested.vmxon_ptr = -1ull;
        vmx->nested.posted_intr_nv = -1;
        vmx->nested.current_vmptr = -1ull;
        vmx->nested.hv_evmcs_vmptr = EVMPTR_INVALID;
Quoted 6 lines
Quoted 2 lines

to avoid issues in the future.

Thanks, Vitaly. And above suggestion sounds reasonable to me. :)

I can add this in V2 later, if no one objects.

Sure, this is rather a separate patch. Thanks!

Signature · 1 line
Vitaly
author
#4 lore raw
Quoted 2 lines · Vitaly Kuznetsov

Sure, this is rather a separate patch. Thanks!

Oh, right. Just sent out a seperate patch. :)

@Paolo Any comment on this one? Thanks!

B.R.
Yu

maintainer Applied
#5 lore raw
In reply to Yu Zhang
Quoted 24 lines · Yu Zhang

"VMXON pointer" is saved in vmx->nested.vmxon_ptr since commit 3573e22cfeca ("KVM: nVMX: additional checks on vmxon region"). Also, handle_vmptrld() & handle_vmclear() now have logic to check the VMCS pointer against the VMXON pointer.

So just remove the obsolete comments of handle_vmon().

---
  arch/x86/kvm/vmx/nested.c | 9 +--------
  1 file changed, 1 insertion(+), 8 deletions(-)
arch/x86/kvm/vmx/nested.c +0 -2
index bc6327950657..90f34f12f883 100644
@@ -4862,14 +4862,7 @@ static int enter_vmx_operation(struct kvm_vcpu *vcpu)
  	return -ENOMEM;
  }
 
-/*
- * Emulate the VMXON instruction.

arch/x86/kvm/vmx/nested.c static int enter_vmx_operation(struct kvm_vcpu *vcpu) · line 4868

Quoted 10 lines
arch/x86/kvm/vmx/nested.c +1 -6
- * Currently, we just remember that VMX is active, and do not save or even
- * inspect the argument to VMXON (the so-called "VMXON pointer") because we
- * do not currently need to store anything in that guest-allocated memory
- * region. Consequently, VMCLEAR and VMPTRLD also do not verify that the their
- * argument is different from the VMXON pointer (which the spec says they do).
- */
+/* Emulate the VMXON instruction. */
  static int handle_vmon(struct kvm_vcpu *vcpu)
  {
  	int ret;

Queued, thanks.

Paolo