A VMX preemption timer value of '0' is guaranteed to cause a VMExit
prior to the CPU executing any instructions in the guest. Use the
preemption timer (if it's supported) to trigger immediate VMExit
in place of the current method of sending a self-IPI. This ensures
that pending VMExit injection to L1 occurs prior to executing any
instructions in the guest (regardless of nesting level).
When deferring VMExit injection, KVM generates an immediate VMExit
from the (possibly nested) guest by sending itself an IPI. Because
hardware interrupts are blocked prior to VMEnter and are unblocked
(in hardware) after VMEnter, this results in taking a VMExit(INTR)
before any guest instruction is executed. But, as this approach
relies on the IPI being received before VMEnter executes, it only
works as intended when KVM is running as L0. Because there are no
architectural guarantees regarding when IPIs are delivered, when
running nested the INTR may "arrive" long after L2 is running e.g.
L0 KVM doesn't force an immediate switch to L1 to deliver an INTR.
For the most part, this unintended delay is not an issue since the
events being injected to L1 also do not have architectural guarantees
regarding their timing. The notable exception is the VMX preemption
timer[1], which is architecturally guaranteed to cause a VMExit prior
to executing any instructions in the guest if the timer value is '0'
at VMEnter. Specifically, the delay in injecting the VMExit causes
the preemption timer KVM unit test to fail when run in a nested guest.
Note: this approach is viable even on CPUs with a broken preemption
timer, as broken in this context only means the timer counts at the
wrong rate. There are no known errata affecting timer value of '0'.
[1] I/O SMIs also have guarantees on when they arrive, but I have
no idea if/how those are emulated in KVM.
Signed-off-by: Sean Christopherson <sean.j.christopherson@xxxxxxxxx>
---
arch/x86/include/asm/kvm_host.h | 1 +
arch/x86/kvm/vmx.c | 22 ++++++++++++++++++----
arch/x86/kvm/x86.c | 5 ++++-
3 files changed, 23 insertions(+), 5 deletions(-)
diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index 00ddb0c9e612..6ae0d3f18ee3 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -1055,6 +1055,7 @@ struct kvm_x86_ops {
bool (*umip_emulated)(void);
int (*check_nested_events)(struct kvm_vcpu *vcpu, bool external_intr);
+ void (*request_immediate_exit)(struct kvm_vcpu *vcpu);
void (*sched_in)(struct kvm_vcpu *kvm, int cpu);
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 5ae46af2077d..2f497fbc62dc 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -1026,6 +1026,7 @@ struct vcpu_vmx {
/* apic deadline value in host tsc */
u64 hv_deadline_tsc;
bool hv_timer_armed;
+ bool req_immediate_exit;
u64 current_tsc_ratio;
@@ -2865,6 +2866,8 @@ static void vmx_prepare_switch_to_guest(struct kvm_vcpu *vcpu)
u16 fs_sel, gs_sel;
int i;
+ vmx->req_immediate_exit = false;
+
if (vmx->loaded_cpu_state)
return;
@@ -7967,6 +7970,9 @@ static __init int hardware_setup(void)
kvm_x86_ops->enable_log_dirty_pt_masked = NULL;
}
+ if (!cpu_has_vmx_preemption_timer())
+ kvm_x86_ops->request_immediate_exit = NULL;
+
if (cpu_has_vmx_preemption_timer() && enable_preemption_timer) {
u64 vmx_msr;
@@ -9209,7 +9215,8 @@ static int handle_pml_full(struct kvm_vcpu *vcpu)
static int handle_preemption_timer(struct kvm_vcpu *vcpu)
{
- kvm_lapic_expired_hv_timer(vcpu);
+ if (!to_vmx(vcpu)->req_immediate_exit)
+ kvm_lapic_expired_hv_timer(vcpu);
return 1;
}
@@ -10606,7 +10613,9 @@ static void vmx_arm_hv_timer(struct kvm_vcpu *vcpu)
u64 tscl;
u32 delta_tsc;
- if (vmx->hv_deadline_tsc != -1) {
+ if (vmx->req_immediate_exit) {
+ vmcs_write32(VMX_PREEMPTION_TIMER_VALUE, 0);
+ } else if (vmx->hv_deadline_tsc != -1) {
tscl = rdtsc();
if (vmx->hv_deadline_tsc > tscl)
/* set_hv_timer ensures the delta fits in 32-bits */
@@ -10614,11 +10623,10 @@ static void vmx_arm_hv_timer(struct kvm_vcpu *vcpu)
cpu_preemption_timer_multi);
else
delta_tsc = 0;
-
vmcs_write32(VMX_PREEMPTION_TIMER_VALUE, delta_tsc);
}
- arm_timer = (vmx->hv_deadline_tsc != -1);
+ arm_timer = (vmx->hv_deadline_tsc != -1) || vmx->req_immediate_exit;
if (arm_timer && !vmx->hv_timer_armed)
vmcs_set_bits(PIN_BASED_VM_EXEC_CONTROL,
PIN_BASED_VMX_PREEMPTION_TIMER);
@@ -12856,6 +12864,11 @@ static int vmx_check_nested_events(struct kvm_vcpu *vcpu, bool external_intr)
return 0;
}
+static void vmx_request_immediate_exit(struct kvm_vcpu *vcpu)
+{
+ to_vmx(vcpu)->req_immediate_exit = true;
+}
+
static u32 vmx_get_preemption_timer_value(struct kvm_vcpu *vcpu)
{
ktime_t remaining =
@@ -14115,6 +14128,7 @@ static struct kvm_x86_ops vmx_x86_ops __ro_after_init = {
.umip_emulated = vmx_umip_emulated,
.check_nested_events = vmx_check_nested_events,
+ .request_immediate_exit = vmx_request_immediate_exit,
.sched_in = vmx_sched_in,
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 14ee9a814888..e2906ec45cb6 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -7544,7 +7544,10 @@ static int vcpu_enter_guest(struct kvm_vcpu *vcpu)
if (req_immediate_exit) {
kvm_make_request(KVM_REQ_EVENT, vcpu);
- smp_send_reschedule(vcpu->cpu);
+ if (kvm_x86_ops->request_immediate_exit)
+ kvm_x86_ops->request_immediate_exit(vcpu);
+ else
+ smp_send_reschedule(vcpu->cpu);
}
trace_kvm_entry(vcpu->vcpu_id);
--
2.18.0
