On Fri, 15 Nov 2019 at 00:34, Sean Christopherson
<sean.j.christopherson@xxxxxxxxx> wrote:
>
> On Thu, Nov 14, 2019 at 04:44:33PM +0100, Paolo Bonzini wrote:
> > On 14/11/19 16:22, Sean Christopherson wrote:
> > >> Instead of a separate vcpu->fast_vmexit, perhaps you can set exit_reason
> > >> to vmx->exit_reason to -1 if the fast path succeeds.
> > >
> > > Actually, rather than make this super special case, what about moving the
> > > handling into vmx_handle_exit_irqoff()? Practically speaking that would
> > > only add ~50 cycles (two VMREADs) relative to the code being run right
> > > after kvm_put_guest_xcr0(). It has the advantage of restoring the host's
> > > hardware breakpoints, preserving a semi-accurate last_guest_tsc, and
> > > running with vcpu->mode set back to OUTSIDE_GUEST_MODE. Hopefully it'd
> > > also be more intuitive for people unfamiliar with the code.
> >
> > Yes, that's a good idea. The expensive bit between handle_exit_irqoff
> > and handle_exit is srcu_read_lock, which has two memory barriers in it.
Moving the handling into vmx_handle_exit_irqoff() can worse ~100
cycles than right after kvm_put_guest_xcr0() in my testing. So, which
one do you prefer?
For moving the handling into vmx_handle_exit_irqoff(), how about the
patch below:
---8<---
>From 1b20b0a80c6da18b721f125cc40f8be5ad31f4b1 Mon Sep 17 00:00:00 2001
From: Wanpeng Li <wanpengli@xxxxxxxxxxx>
Date: Fri, 15 Nov 2019 10:18:09 +0800
Subject: [PATCH v2] KVM: VMX: FIXED+PHYSICAL mode single target IPI fastpath
Signed-off-by: Wanpeng Li <wanpengli@xxxxxxxxxxx>
---
arch/x86/include/asm/kvm_host.h | 4 ++--
arch/x86/include/uapi/asm/vmx.h | 1 +
arch/x86/kvm/svm.c | 4 ++--
arch/x86/kvm/vmx/vmx.c | 40 +++++++++++++++++++++++++++++++++++++---
arch/x86/kvm/x86.c | 5 +++--
5 files changed, 45 insertions(+), 9 deletions(-)
diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index 24d6598..3604f3a 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -1067,7 +1067,7 @@ struct kvm_x86_ops {
void (*tlb_flush_gva)(struct kvm_vcpu *vcpu, gva_t addr);
void (*run)(struct kvm_vcpu *vcpu);
- int (*handle_exit)(struct kvm_vcpu *vcpu);
+ int (*handle_exit)(struct kvm_vcpu *vcpu, u32 *vcpu_exit_reason);
int (*skip_emulated_instruction)(struct kvm_vcpu *vcpu);
void (*set_interrupt_shadow)(struct kvm_vcpu *vcpu, int mask);
u32 (*get_interrupt_shadow)(struct kvm_vcpu *vcpu);
@@ -1117,7 +1117,7 @@ struct kvm_x86_ops {
int (*check_intercept)(struct kvm_vcpu *vcpu,
struct x86_instruction_info *info,
enum x86_intercept_stage stage);
- void (*handle_exit_irqoff)(struct kvm_vcpu *vcpu);
+ void (*handle_exit_irqoff)(struct kvm_vcpu *vcpu, u32 *vcpu_exit_reason);
bool (*mpx_supported)(void);
bool (*xsaves_supported)(void);
bool (*umip_emulated)(void);
diff --git a/arch/x86/include/uapi/asm/vmx.h b/arch/x86/include/uapi/asm/vmx.h
index 3eb8411..b33c6e1 100644
--- a/arch/x86/include/uapi/asm/vmx.h
+++ b/arch/x86/include/uapi/asm/vmx.h
@@ -88,6 +88,7 @@
#define EXIT_REASON_XRSTORS 64
#define EXIT_REASON_UMWAIT 67
#define EXIT_REASON_TPAUSE 68
+#define EXIT_REASON_NEED_SKIP_EMULATED_INSN -1
#define VMX_EXIT_REASONS \
{ EXIT_REASON_EXCEPTION_NMI, "EXCEPTION_NMI" }, \
diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
index c5673bd..3bb0661 100644
--- a/arch/x86/kvm/svm.c
+++ b/arch/x86/kvm/svm.c
@@ -4927,7 +4927,7 @@ static void svm_get_exit_info(struct kvm_vcpu
*vcpu, u64 *info1, u64 *info2)
*info2 = control->exit_info_2;
}
-static int handle_exit(struct kvm_vcpu *vcpu)
+static int handle_exit(struct kvm_vcpu *vcpu, u32 *vcpu_exit_reason)
{
struct vcpu_svm *svm = to_svm(vcpu);
struct kvm_run *kvm_run = vcpu->run;
@@ -6171,7 +6171,7 @@ static int svm_check_intercept(struct kvm_vcpu *vcpu,
return ret;
}
-static void svm_handle_exit_irqoff(struct kvm_vcpu *vcpu)
+static void svm_handle_exit_irqoff(struct kvm_vcpu *vcpu, u32
*vcpu_exit_reason)
{
}
diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
index 5d21a4a..073fe1f 100644
--- a/arch/x86/kvm/vmx/vmx.c
+++ b/arch/x86/kvm/vmx/vmx.c
@@ -5838,7 +5838,7 @@ void dump_vmcs(void)
* The guest has exited. See if we can fix it or if we need userspace
* assistance.
*/
-static int vmx_handle_exit(struct kvm_vcpu *vcpu)
+static int vmx_handle_exit(struct kvm_vcpu *vcpu, u32 *vcpu_exit_reason)
{
struct vcpu_vmx *vmx = to_vmx(vcpu);
u32 exit_reason = vmx->exit_reason;
@@ -5924,7 +5924,10 @@ static int vmx_handle_exit(struct kvm_vcpu *vcpu)
}
}
- if (exit_reason < kvm_vmx_max_exit_handlers
+ if (*vcpu_exit_reason == EXIT_REASON_NEED_SKIP_EMULATED_INSN) {
+ kvm_skip_emulated_instruction(vcpu);
+ return 1;
+ } else if (exit_reason < kvm_vmx_max_exit_handlers
&& kvm_vmx_exit_handlers[exit_reason])
return kvm_vmx_exit_handlers[exit_reason](vcpu);
else {
@@ -6255,7 +6258,36 @@ static void
handle_external_interrupt_irqoff(struct kvm_vcpu *vcpu)
}
STACK_FRAME_NON_STANDARD(handle_external_interrupt_irqoff);
-static void vmx_handle_exit_irqoff(struct kvm_vcpu *vcpu)
+static u32 handle_ipi_fastpath(struct kvm_vcpu *vcpu)
+{
+ u32 index;
+ u64 data;
+ int ret = 0;
+
+ if (lapic_in_kernel(vcpu) && apic_x2apic_mode(vcpu->arch.apic)) {
+ /*
+ * fastpath to IPI target, FIXED+PHYSICAL which is popular
+ */
+ index = kvm_rcx_read(vcpu);
+ data = kvm_read_edx_eax(vcpu);
+
+ if (((index - APIC_BASE_MSR) << 4 == APIC_ICR) &&
+ ((data & KVM_APIC_DEST_MASK) == APIC_DEST_PHYSICAL) &&
+ ((data & APIC_MODE_MASK) == APIC_DM_FIXED)) {
+
+ trace_kvm_msr_write(index, data);
+ kvm_lapic_set_reg(vcpu->arch.apic, APIC_ICR2, (u32)(data >> 32));
+ ret = kvm_lapic_reg_write(vcpu->arch.apic, APIC_ICR, (u32)data);
+
+ if (ret == 0)
+ return EXIT_REASON_NEED_SKIP_EMULATED_INSN;
+ }
+ }
+
+ return ret;
+}
+
+static void vmx_handle_exit_irqoff(struct kvm_vcpu *vcpu, u32 *exit_reason)
{
struct vcpu_vmx *vmx = to_vmx(vcpu);
@@ -6263,6 +6295,8 @@ static void vmx_handle_exit_irqoff(struct kvm_vcpu *vcpu)
handle_external_interrupt_irqoff(vcpu);
else if (vmx->exit_reason == EXIT_REASON_EXCEPTION_NMI)
handle_exception_nmi_irqoff(vmx);
+ else if (vmx->exit_reason == EXIT_REASON_MSR_WRITE)
+ *exit_reason = handle_ipi_fastpath(vcpu);
}
static bool vmx_has_emulated_msr(int index)
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index ff395f8..130576b 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -7931,6 +7931,7 @@ EXPORT_SYMBOL_GPL(__kvm_request_immediate_exit);
static int vcpu_enter_guest(struct kvm_vcpu *vcpu)
{
int r;
+ u32 exit_reason = 0;
bool req_int_win =
dm_request_for_irq_injection(vcpu) &&
kvm_cpu_accept_dm_intr(vcpu);
@@ -8180,7 +8181,7 @@ static int vcpu_enter_guest(struct kvm_vcpu *vcpu)
vcpu->mode = OUTSIDE_GUEST_MODE;
smp_wmb();
- kvm_x86_ops->handle_exit_irqoff(vcpu);
+ kvm_x86_ops->handle_exit_irqoff(vcpu, &exit_reason);
/*
* Consume any pending interrupts, including the possible source of
@@ -8224,7 +8225,7 @@ static int vcpu_enter_guest(struct kvm_vcpu *vcpu)
kvm_lapic_sync_from_vapic(vcpu);
vcpu->arch.gpa_available = false;
- r = kvm_x86_ops->handle_exit(vcpu);
+ r = kvm_x86_ops->handle_exit(vcpu, &exit_reason);
return r;
cancel_injection:
--
2.7.4
