[PATCH v2 3/3] x86/kvm: Fix vtime accounting

[Wanpeng Li]

From: Wanpeng Li <wanpengli@xxxxxxxxxxx>

The bugzilla https://bugzilla.kernel.org/show_bug.cgi?id=209831
reported that the guest time remains 0 when running a while true
loop in the guest.

The commit 87fa7f3e98a131 ("x86/kvm: Move context tracking where it
belongs") moves guest_exit_irqoff() close to vmexit breaks the
tick-based time accouting when the ticks that happen after IRQs are
disabled are incorrectly accounted to the host/system time. This is
because we exit the guest state too early.

Keep context tracking around the actual vmentry/exit code, the time 
accounting logic is separated out by preposed patch from 
guest_enter/exit_irqoff(). Keep vtime-based time accounting around 
the actual vmentry/exit code when vtime_accounting_enabled_this_cpu() 
is true, leave PF_VCPU clearing after handle_exit_irqoff() and explicit 
IRQ window for tick-based time accouting.

Fixes: 87fa7f3e98a131 ("x86/kvm: Move context tracking where it belongs")
Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Cc: Sean Christopherson <seanjc@xxxxxxxxxx>
Cc: Michael Tokarev <mjt@xxxxxxxxxx>
Cc: stable@xxxxxxxxxxxxxxx#v5.9-rc1+
Suggested-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Signed-off-by: Wanpeng Li <wanpengli@xxxxxxxxxxx>
---
 arch/x86/kvm/svm/svm.c | 6 ++++--
 arch/x86/kvm/vmx/vmx.c | 6 ++++--
 arch/x86/kvm/x86.c     | 1 +
 3 files changed, 9 insertions(+), 4 deletions(-)

diff --git a/arch/x86/kvm/svm/svm.c b/arch/x86/kvm/svm/svm.c
index 48b396f3..2a4e284 100644
--- a/arch/x86/kvm/svm/svm.c
+++ b/arch/x86/kvm/svm/svm.c
@@ -3726,11 +3726,12 @@ static noinstr void svm_vcpu_enter_exit(struct kvm_vcpu *vcpu)
 	 * accordingly.
 	 */
 	instrumentation_begin();
+	vtime_account_guest_enter();
 	trace_hardirqs_on_prepare();
 	lockdep_hardirqs_on_prepare(CALLER_ADDR0);
 	instrumentation_end();
 
-	guest_enter_irqoff();
+	context_guest_enter_irqoff();
 	lockdep_hardirqs_on(CALLER_ADDR0);
 
 	if (sev_es_guest(vcpu->kvm)) {
@@ -3758,10 +3759,11 @@ static noinstr void svm_vcpu_enter_exit(struct kvm_vcpu *vcpu)
 	 * into world and some more.
 	 */
 	lockdep_hardirqs_off(CALLER_ADDR0);
-	guest_exit_irqoff();
+	context_guest_exit_irqoff();
 
 	instrumentation_begin();
 	trace_hardirqs_off_finish();
+	__vtime_account_guest_exit();
 	instrumentation_end();
 }
 
diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
index c05e6e2..23be956 100644
--- a/arch/x86/kvm/vmx/vmx.c
+++ b/arch/x86/kvm/vmx/vmx.c
@@ -6613,11 +6613,12 @@ static noinstr void vmx_vcpu_enter_exit(struct kvm_vcpu *vcpu,
 	 * accordingly.
 	 */
 	instrumentation_begin();
+	vtime_account_guest_enter();
 	trace_hardirqs_on_prepare();
 	lockdep_hardirqs_on_prepare(CALLER_ADDR0);
 	instrumentation_end();
 
-	guest_enter_irqoff();
+	context_guest_enter_irqoff();
 	lockdep_hardirqs_on(CALLER_ADDR0);
 
 	/* L1D Flush includes CPU buffer clear to mitigate MDS */
@@ -6647,10 +6648,11 @@ static noinstr void vmx_vcpu_enter_exit(struct kvm_vcpu *vcpu,
 	 * into world and some more.
 	 */
 	lockdep_hardirqs_off(CALLER_ADDR0);
-	guest_exit_irqoff();
+	context_guest_exit_irqoff();
 
 	instrumentation_begin();
 	trace_hardirqs_off_finish();
+	__vtime_account_guest_exit();
 	instrumentation_end();
 }
 
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index ce9a1d2..0d2dd3f 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -9245,6 +9245,7 @@ static int vcpu_enter_guest(struct kvm_vcpu *vcpu)
 	++vcpu->stat.exits;
 	local_irq_disable();
 	kvm_after_interrupt(vcpu);
+	vcpu_account_guest_exit();
 
 	if (lapic_in_kernel(vcpu)) {
 		s64 delta = vcpu->arch.apic->lapic_timer.advance_expire_delta;
-- 
2.7.4