KVM: x86: do not attempt TSC synchronization on guest writes

KVM special-cases writes to MSR_IA32_TSC so that all CPUs have
the same base for the TSC.  This logic is complicated, and we
do not want it to have any effect once the VM is started.

In particular, if any guest started to synchronize its TSCs
with writes to MSR_IA32_TSC rather than MSR_IA32_TSC_ADJUST,
the additional effect of kvm_write_tsc code would be uncharted
territory.

Therefore, this patch makes writes to MSR_IA32_TSC behave
essentially the same as writes to MSR_IA32_TSC_ADJUST when
they come from the guest.  A new selftest (which passes
both before and after the patch) checks the current semantics
of writes to MSR_IA32_TSC and MSR_IA32_TSC_ADJUST originating
from both the host and the guest.

Upcoming work to remove the special side effects
of host-initiated writes to MSR_IA32_TSC and MSR_IA32_TSC_ADJUST
will be able to build onto this test, adjusting the host side
to use the new APIs and achieve the same effect.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

1 files changed, 10 insertions, 20 deletions

diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 411f6103532b..c4015a43cc8a 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -2107,12 +2107,6 @@ static void kvm_track_tsc_matching(struct kvm_vcpu *vcpu)
 #endif
 }
 
-static void update_ia32_tsc_adjust_msr(struct kvm_vcpu *vcpu, s64 offset)
-{
-	u64 curr_offset = vcpu->arch.l1_tsc_offset;
-	vcpu->arch.ia32_tsc_adjust_msr += offset - curr_offset;
-}
-
 /*
  * Multiply tsc by a fixed point number represented by ratio.
  *
@@ -2174,14 +2168,13 @@ static inline bool kvm_check_tsc_unstable(void)
 	return check_tsc_unstable();
 }
 
-void kvm_write_tsc(struct kvm_vcpu *vcpu, struct msr_data *msr)
+static void kvm_synchronize_tsc(struct kvm_vcpu *vcpu, u64 data)
 {
 	struct kvm *kvm = vcpu->kvm;
 	u64 offset, ns, elapsed;
 	unsigned long flags;
 	bool matched;
 	bool already_matched;
-	u64 data = msr->data;
 	bool synchronizing = false;
 
 	raw_spin_lock_irqsave(&kvm->arch.tsc_write_lock, flags);
@@ -2190,7 +2183,7 @@ void kvm_write_tsc(struct kvm_vcpu *vcpu, struct msr_data *msr)
 	elapsed = ns - kvm->arch.last_tsc_nsec;
 
 	if (vcpu->arch.virtual_tsc_khz) {
-		if (data == 0 && msr->host_initiated) {
+		if (data == 0) {
 			/*
 			 * detection of vcpu initialization -- need to sync
 			 * with other vCPUs. This particularly helps to keep
@@ -2260,9 +2253,6 @@ void kvm_write_tsc(struct kvm_vcpu *vcpu, struct msr_data *msr)
 	vcpu->arch.this_tsc_nsec = kvm->arch.cur_tsc_nsec;
 	vcpu->arch.this_tsc_write = kvm->arch.cur_tsc_write;
 
-	if (!msr->host_initiated && guest_cpuid_has(vcpu, X86_FEATURE_TSC_ADJUST))
-		update_ia32_tsc_adjust_msr(vcpu, offset);
-
 	kvm_vcpu_write_tsc_offset(vcpu, offset);
 	raw_spin_unlock_irqrestore(&kvm->arch.tsc_write_lock, flags);
 
@@ -2277,8 +2267,6 @@ void kvm_write_tsc(struct kvm_vcpu *vcpu, struct msr_data *msr)
 	spin_unlock(&kvm->arch.pvclock_gtod_sync_lock);
 }
 
-EXPORT_SYMBOL_GPL(kvm_write_tsc);
-
 static inline void adjust_tsc_offset_guest(struct kvm_vcpu *vcpu,
 					   s64 adjustment)
 {
@@ -3073,7 +3061,13 @@ int kvm_set_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
 		vcpu->arch.msr_ia32_power_ctl = data;
 		break;
 	case MSR_IA32_TSC:
-		kvm_write_tsc(vcpu, msr_info);
+		if (msr_info->host_initiated) {
+			kvm_synchronize_tsc(vcpu, data);
+		} else {
+			u64 adj = kvm_compute_tsc_offset(vcpu, data) - vcpu->arch.l1_tsc_offset;
+			adjust_tsc_offset_guest(vcpu, adj);
+			vcpu->arch.ia32_tsc_adjust_msr += adj;
+		}
 		break;
 	case MSR_IA32_XSS:
 		if (!msr_info->host_initiated &&
@@ -9839,7 +9833,6 @@ fail_mmu_destroy:
 
 void kvm_arch_vcpu_postcreate(struct kvm_vcpu *vcpu)
 {
-	struct msr_data msr;
 	struct kvm *kvm = vcpu->kvm;
 
 	kvm_hv_vcpu_postcreate(vcpu);
@@ -9847,10 +9840,7 @@ void kvm_arch_vcpu_postcreate(struct kvm_vcpu *vcpu)
 	if (mutex_lock_killable(&vcpu->mutex))
 		return;
 	vcpu_load(vcpu);
-	msr.data = 0x0;
-	msr.index = MSR_IA32_TSC;
-	msr.host_initiated = true;
-	kvm_write_tsc(vcpu, &msr);
+	kvm_synchronize_tsc(vcpu, 0);
 	vcpu_put(vcpu);
 
 	/* poll control enabled by default */