Defer fpu deactivation as much as possible - if the guest fpu is loaded, keep
it loaded until the next heavyweight exit (where we are forced to unload it).
This reduces unnecessary exits.
We also defer fpu activation on clts; while clts signals the intent to use the
fpu, we can't be sure the guest will actually use it.
Signed-off-by: Avi Kivity <avi@xxxxxxxxxx>
---
arch/x86/include/asm/kvm_host.h | 1 +
arch/x86/kvm/svm.c | 34 +++++++++++++++++++---------------
arch/x86/kvm/vmx.c | 13 +------------
arch/x86/kvm/x86.c | 6 +++++-
4 files changed, 26 insertions(+), 28 deletions(-)
diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index 627d50b..b3b3ce1 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -498,6 +498,7 @@ struct kvm_x86_ops {
void (*cache_reg)(struct kvm_vcpu *vcpu, enum kvm_reg reg);
unsigned long (*get_rflags)(struct kvm_vcpu *vcpu);
void (*set_rflags)(struct kvm_vcpu *vcpu, unsigned long rflags);
+ void (*fpu_deactivate)(struct kvm_vcpu *vcpu);
void (*tlb_flush)(struct kvm_vcpu *vcpu);
diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
index b2a7155..2a3890f 100644
--- a/arch/x86/kvm/svm.c
+++ b/arch/x86/kvm/svm.c
@@ -984,17 +984,8 @@ static void svm_set_cr0(struct kvm_vcpu *vcpu, unsigned long cr0)
if (npt_enabled)
goto set;
- if (kvm_read_cr0_bits(vcpu, X86_CR0_TS) && !(cr0 & X86_CR0_TS)) {
- svm->vmcb->control.intercept_exceptions &= ~(1 << NM_VECTOR);
- vcpu->fpu_active = 1;
- }
-
vcpu->arch.cr0 = cr0;
cr0 |= X86_CR0_PG | X86_CR0_WP;
- if (!vcpu->fpu_active) {
- svm->vmcb->control.intercept_exceptions |= (1 << NM_VECTOR);
- cr0 |= X86_CR0_TS;
- }
set:
/*
* re-enable caching here because the QEMU bios
@@ -1250,6 +1241,8 @@ static int nm_interception(struct vcpu_svm *svm)
svm->vmcb->control.intercept_exceptions &= ~(1 << NM_VECTOR);
if (!kvm_read_cr0_bits(&svm->vcpu, X86_CR0_TS))
svm->vmcb->save.cr0 &= ~X86_CR0_TS;
+ else
+ svm->vmcb->save.cr0 |= X86_CR0_TS;
svm->vcpu.fpu_active = 1;
return 1;
@@ -2586,6 +2579,8 @@ static void svm_flush_tlb(struct kvm_vcpu *vcpu)
static void svm_prepare_guest_switch(struct kvm_vcpu *vcpu)
{
+ if (npt_enabled)
+ vcpu->fpu_active = 1;
}
static inline void sync_cr8_to_lapic(struct kvm_vcpu *vcpu)
@@ -2805,12 +2800,6 @@ static void svm_set_cr3(struct kvm_vcpu *vcpu, unsigned long root)
svm->vmcb->save.cr3 = root;
force_new_asid(vcpu);
-
- if (vcpu->fpu_active) {
- svm->vmcb->control.intercept_exceptions |= (1 << NM_VECTOR);
- svm->vmcb->save.cr0 |= X86_CR0_TS;
- vcpu->fpu_active = 0;
- }
}
static int is_disabled(void)
@@ -2926,6 +2915,20 @@ static bool svm_rdtscp_supported(void)
return false;
}
+static void svm_fpu_deactivate(struct kvm_vcpu *vcpu)
+{
+ struct vcpu_svm *svm = to_svm(vcpu);
+
+ if (npt_enabled) {
+ /* hack: npt requires active fpu at this time */
+ vcpu->fpu_active = 1;
+ return;
+ }
+
+ svm->vmcb->control.intercept_exceptions |= 1 << NM_VECTOR;
+ svm->vmcb->save.cr0 |= X86_CR0_TS;
+}
+
static struct kvm_x86_ops svm_x86_ops = {
.cpu_has_kvm_support = has_svm,
.disabled_by_bios = is_disabled,
@@ -2967,6 +2970,7 @@ static struct kvm_x86_ops svm_x86_ops = {
.cache_reg = svm_cache_reg,
.get_rflags = svm_get_rflags,
.set_rflags = svm_set_rflags,
+ .fpu_deactivate = svm_fpu_deactivate,
.tlb_flush = svm_flush_tlb,
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 0baf62f..7e0b45e 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -801,9 +801,6 @@ static void vmx_fpu_activate(struct kvm_vcpu *vcpu)
static void vmx_fpu_deactivate(struct kvm_vcpu *vcpu)
{
- if (!vcpu->fpu_active)
- return;
- vcpu->fpu_active = 0;
vmcs_set_bits(GUEST_CR0, X86_CR0_TS);
update_exception_bitmap(vcpu);
}
@@ -1727,8 +1724,6 @@ static void vmx_set_cr0(struct kvm_vcpu *vcpu, unsigned long cr0)
else
hw_cr0 = (cr0 & ~KVM_GUEST_CR0_MASK) | KVM_VM_CR0_ALWAYS_ON;
- vmx_fpu_deactivate(vcpu);
-
if (vmx->rmode.vm86_active && (cr0 & X86_CR0_PE))
enter_pmode(vcpu);
@@ -1750,9 +1745,6 @@ static void vmx_set_cr0(struct kvm_vcpu *vcpu, unsigned long cr0)
vmcs_writel(CR0_READ_SHADOW, cr0);
vmcs_writel(GUEST_CR0, hw_cr0);
vcpu->arch.cr0 = cr0;
-
- if (!(cr0 & X86_CR0_TS) || !(cr0 & X86_CR0_PE))
- vmx_fpu_activate(vcpu);
}
static u64 construct_eptp(unsigned long root_hpa)
@@ -1783,8 +1775,6 @@ static void vmx_set_cr3(struct kvm_vcpu *vcpu, unsigned long cr3)
vmx_flush_tlb(vcpu);
vmcs_writel(GUEST_CR3, guest_cr3);
- if (kvm_read_cr0_bits(vcpu, X86_CR0_PE))
- vmx_fpu_deactivate(vcpu);
}
static void vmx_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4)
@@ -2992,11 +2982,9 @@ static int handle_cr(struct kvm_vcpu *vcpu)
};
break;
case 2: /* clts */
- vmx_fpu_deactivate(vcpu);
vcpu->arch.cr0 &= ~X86_CR0_TS;
vmcs_writel(CR0_READ_SHADOW, kvm_read_cr0(vcpu));
trace_kvm_cr_write(0, kvm_read_cr0(vcpu));
- vmx_fpu_activate(vcpu);
skip_emulated_instruction(vcpu);
return 1;
case 1: /*mov from cr*/
@@ -4093,6 +4081,7 @@ static struct kvm_x86_ops vmx_x86_ops = {
.cache_reg = vmx_cache_reg,
.get_rflags = vmx_get_rflags,
.set_rflags = vmx_set_rflags,
+ .fpu_deactivate = vmx_fpu_deactivate,
.tlb_flush = vmx_flush_tlb,
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 78c4397..28c2020 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -1509,8 +1509,8 @@ void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
void kvm_arch_vcpu_put(struct kvm_vcpu *vcpu)
{
- kvm_x86_ops->vcpu_put(vcpu);
kvm_put_guest_fpu(vcpu);
+ kvm_x86_ops->vcpu_put(vcpu);
}
static int is_efer_nx(void)
@@ -4988,6 +4988,10 @@ void kvm_put_guest_fpu(struct kvm_vcpu *vcpu)
return;
vcpu->guest_fpu_loaded = 0;
+ if (vcpu->fpu_active) {
+ vcpu->fpu_active = 0;
+ kvm_x86_ops->fpu_deactivate(vcpu);
+ }
kvm_fx_save(&vcpu->arch.guest_fx_image);
kvm_fx_restore(&vcpu->arch.host_fx_image);
++vcpu->stat.fpu_reload;
--
1.6.5.3
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
