[PATCH v2 0/2] minor cleanups on efer emulation

[PATCH v2 0/2] minor cleanups on efer emulation

[Zhenzhong Duan]

These two patches remove some redundant code related to mode switch
and EFER emulation.

Sanity tested with kernel in L0, L1 and L2 all patched.

v2: Split to two patches and use comments from Sean to explain why
it's secure to remove them, suggested by Sean.

Zhenzhong Duan (2):
  KVM: x86: Remove unnecessory assignment to uret->data
  KVM: x86: Remove redundant vm_entry_controls_clearbit() call

 arch/x86/kvm/vmx/vmx.c | 15 +++++----------
 1 file changed, 5 insertions(+), 10 deletions(-)

-- 
2.25.1

[PATCH v2 1/2] KVM: x86: Remove unnecessory assignment to uret->data

[Zhenzhong Duan]

While update_transition_efer() doesn't unconditionally set uret->data,
which on the surface makes this look suspect, but it's safe because
uret->data is consumed if and only if uret->load_into_hardware is
true, and it's (a) set to false if uret->data isn't updated and
(b) uret->data is guaranteed to be updated before it's set to true.

Drop the local "msr" and use "vmx" directly instead of redoing to_vmx().

Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
---
 arch/x86/kvm/vmx/vmx.c | 14 +++++---------
 1 file changed, 5 insertions(+), 9 deletions(-)

diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
index b730d799c26e..cadb3769031c 100644
--- a/arch/x86/kvm/vmx/vmx.c
+++ b/arch/x86/kvm/vmx/vmx.c
@@ -2871,21 +2871,17 @@ static void enter_rmode(struct kvm_vcpu *vcpu)
 int vmx_set_efer(struct kvm_vcpu *vcpu, u64 efer)
 {
 	struct vcpu_vmx *vmx = to_vmx(vcpu);
-	struct vmx_uret_msr *msr = vmx_find_uret_msr(vmx, MSR_EFER);
 
 	/* Nothing to do if hardware doesn't support EFER. */
-	if (!msr)
+	if (!vmx_find_uret_msr(vmx, MSR_EFER))
 		return 0;
 
 	vcpu->arch.efer = efer;
-	if (efer & EFER_LMA) {
-		vm_entry_controls_setbit(to_vmx(vcpu), VM_ENTRY_IA32E_MODE);
-		msr->data = efer;
-	} else {
-		vm_entry_controls_clearbit(to_vmx(vcpu), VM_ENTRY_IA32E_MODE);
+	if (efer & EFER_LMA)
+		vm_entry_controls_setbit(vmx, VM_ENTRY_IA32E_MODE);
+	else
+		vm_entry_controls_clearbit(vmx, VM_ENTRY_IA32E_MODE);
 
-		msr->data = efer & ~EFER_LME;
-	}
 	vmx_setup_uret_msrs(vmx);
 	return 0;
 }
-- 
2.25.1

[PATCH v2 2/2] KVM: x86: Remove redundant vm_entry_controls_clearbit() call

[Zhenzhong Duan]

When emulating exit from long mode, EFER_LMA is cleared which lead to
efer writing emulation, which will unset VM_ENTRY_IA32E_MODE control
bit as requested by SDM. So no need to unset VM_ENTRY_IA32E_MODE again
in exit_lmode() explicitly.

In fact benefited from shadow controls mechanism, this change doesn't
eliminate vmread or vmwrite.

In case EFER isn't supported by hardware, long mode isn't supported,
so this will no break.

Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
---
 arch/x86/kvm/vmx/vmx.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
index cadb3769031c..70717f56a2a2 100644
--- a/arch/x86/kvm/vmx/vmx.c
+++ b/arch/x86/kvm/vmx/vmx.c
@@ -2907,7 +2907,6 @@ static void enter_lmode(struct kvm_vcpu *vcpu)
 
 static void exit_lmode(struct kvm_vcpu *vcpu)
 {
-	vm_entry_controls_clearbit(to_vmx(vcpu), VM_ENTRY_IA32E_MODE);
 	vmx_set_efer(vcpu, vcpu->arch.efer & ~EFER_LMA);
 }
 
-- 
2.25.1

Re: [PATCH v2 0/2] minor cleanups on efer emulation

[Paolo Bonzini]

Queued, thanks.

Paolo