KVM: x86: Add kvm_skip_emulated_instruction and use it.

kvm_skip_emulated_instruction calls both
kvm_x86_ops->skip_emulated_instruction and kvm_vcpu_check_singlestep,
skipping the emulated instruction and generating a trap if necessary.

Replacing skip_emulated_instruction calls with
kvm_skip_emulated_instruction is straightforward, except for:

- ICEBP, which is already inside a trap, so avoid triggering another trap.
- Instructions that can trigger exits to userspace, such as the IO insns,
  MOVs to CR8, and HALT. If kvm_skip_emulated_instruction does trigger a
  KVM_GUESTDBG_SINGLESTEP exit, and the handling code for
  IN/OUT/MOV CR8/HALT also triggers an exit to userspace, the latter will
  take precedence. The singlestep will be triggered again on the next
  instruction, which is the current behavior.
- Task switch instructions which would require additional handling (e.g.
  the task switch bit) and are instead left alone.
- Cases where VMLAUNCH/VMRESUME do not proceed to the next instruction,
  which do not trigger singlestep traps as mentioned previously.

Signed-off-by: Kyle Huey <khuey@kylehuey.com>
Signed-off-by: Radim Krčmář <rkrcmar@redhat.com>

5 files changed, 103 insertions, 124 deletions

diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index 80bad5c372bf..8d1587092851 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -1368,7 +1368,8 @@ void kvm_arch_async_page_ready(struct kvm_vcpu *vcpu,
 bool kvm_arch_can_inject_async_page_present(struct kvm_vcpu *vcpu);
 extern bool kvm_find_async_pf_gfn(struct kvm_vcpu *vcpu, gfn_t gfn);
 
-void kvm_complete_insn_gp(struct kvm_vcpu *vcpu, int err);
+int kvm_skip_emulated_instruction(struct kvm_vcpu *vcpu);
+int kvm_complete_insn_gp(struct kvm_vcpu *vcpu, int err);
 
 int kvm_is_in_guest(void);
 
diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c
index 07cc62955520..dc2685e3f8ea 100644
--- a/arch/x86/kvm/cpuid.c
+++ b/arch/x86/kvm/cpuid.c
@@ -890,7 +890,6 @@ int kvm_emulate_cpuid(struct kvm_vcpu *vcpu)
         kvm_register_write(vcpu, VCPU_REGS_RBX, ebx);
         kvm_register_write(vcpu, VCPU_REGS_RCX, ecx);
         kvm_register_write(vcpu, VCPU_REGS_RDX, edx);
-        kvm_x86_ops->skip_emulated_instruction(vcpu);
-        return 1;
+        return kvm_skip_emulated_instruction(vcpu);
 }
 EXPORT_SYMBOL_GPL(kvm_emulate_cpuid);
diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
index 5bdffcd781f5..08a4d3ab3455 100644
--- a/arch/x86/kvm/svm.c
+++ b/arch/x86/kvm/svm.c
@@ -3151,8 +3151,7 @@ static int skinit_interception(struct vcpu_svm *svm)
 
 static int wbinvd_interception(struct vcpu_svm *svm)
 {
-        kvm_emulate_wbinvd(&svm->vcpu);
-        return 1;
+        return kvm_emulate_wbinvd(&svm->vcpu);
 }
 
 static int xsetbv_interception(struct vcpu_svm *svm)
@@ -3275,9 +3274,7 @@ static int rdpmc_interception(struct vcpu_svm *svm)
                 return emulate_on_interception(svm);
 
         err = kvm_rdpmc(&svm->vcpu);
-        kvm_complete_insn_gp(&svm->vcpu, err);
-
-        return 1;
+        return kvm_complete_insn_gp(&svm->vcpu, err);
 }
 
 static bool check_selective_cr0_intercepted(struct vcpu_svm *svm,
@@ -3374,9 +3371,7 @@ static int cr_interception(struct vcpu_svm *svm)
                 }
                 kvm_register_write(&svm->vcpu, reg, val);
         }
-        kvm_complete_insn_gp(&svm->vcpu, err);
-
-        return 1;
+        return kvm_complete_insn_gp(&svm->vcpu, err);
 }
 
 static int dr_interception(struct vcpu_svm *svm)
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index f4f6304f9583..16a144d22033 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -5556,7 +5556,7 @@ static int handle_triple_fault(struct kvm_vcpu *vcpu)
 static int handle_io(struct kvm_vcpu *vcpu)
 {
         unsigned long exit_qualification;
-        int size, in, string;
+        int size, in, string, ret;
         unsigned port;
 
         exit_qualification = vmcs_readl(EXIT_QUALIFICATION);
@@ -5570,9 +5570,14 @@ static int handle_io(struct kvm_vcpu *vcpu)
 
         port = exit_qualification >> 16;
         size = (exit_qualification & 7) + 1;
-        skip_emulated_instruction(vcpu);
 
-        return kvm_fast_pio_out(vcpu, size, port);
+        ret = kvm_skip_emulated_instruction(vcpu);
+
+        /*
+         * TODO: we might be squashing a KVM_GUESTDBG_SINGLESTEP-triggered
+         * KVM_EXIT_DEBUG here.
+         */
+        return kvm_fast_pio_out(vcpu, size, port) && ret;
 }
 
 static void
@@ -5670,6 +5675,7 @@ static int handle_cr(struct kvm_vcpu *vcpu)
         int cr;
         int reg;
         int err;
+        int ret;
 
         exit_qualification = vmcs_readl(EXIT_QUALIFICATION);
         cr = exit_qualification & 15;
@@ -5681,25 +5687,27 @@ static int handle_cr(struct kvm_vcpu *vcpu)
                 switch (cr) {
                 case 0:
                         err = handle_set_cr0(vcpu, val);
-                        kvm_complete_insn_gp(vcpu, err);
-                        return 1;
+                        return kvm_complete_insn_gp(vcpu, err);
                 case 3:
                         err = kvm_set_cr3(vcpu, val);
-                        kvm_complete_insn_gp(vcpu, err);
-                        return 1;
+                        return kvm_complete_insn_gp(vcpu, err);
                 case 4:
                         err = handle_set_cr4(vcpu, val);
-                        kvm_complete_insn_gp(vcpu, err);
-                        return 1;
+                        return kvm_complete_insn_gp(vcpu, err);
                 case 8: {
                                 u8 cr8_prev = kvm_get_cr8(vcpu);
                                 u8 cr8 = (u8)val;
                                 err = kvm_set_cr8(vcpu, cr8);
-                                kvm_complete_insn_gp(vcpu, err);
+                                ret = kvm_complete_insn_gp(vcpu, err);
                                 if (lapic_in_kernel(vcpu))
-                                        return 1;
+                                        return ret;
                                 if (cr8_prev <= cr8)
-                                        return 1;
+                                        return ret;
+                                /*
+                                 * TODO: we might be squashing a
+                                 * KVM_GUESTDBG_SINGLESTEP-triggered
+                                 * KVM_EXIT_DEBUG here.
+                                 */
                                 vcpu->run->exit_reason = KVM_EXIT_SET_TPR;
                                 return 0;
                         }
@@ -5709,22 +5717,19 @@ static int handle_cr(struct kvm_vcpu *vcpu)
                 handle_clts(vcpu);
                 trace_kvm_cr_write(0, kvm_read_cr0(vcpu));
                 vmx_fpu_activate(vcpu);
-                skip_emulated_instruction(vcpu);
-                return 1;
+                return kvm_skip_emulated_instruction(vcpu);
         case 1: /*mov from cr*/
                 switch (cr) {
                 case 3:
                         val = kvm_read_cr3(vcpu);
                         kvm_register_write(vcpu, reg, val);
                         trace_kvm_cr_read(cr, val);
-                        skip_emulated_instruction(vcpu);
-                        return 1;
+                        return kvm_skip_emulated_instruction(vcpu);
                 case 8:
                         val = kvm_get_cr8(vcpu);
                         kvm_register_write(vcpu, reg, val);
                         trace_kvm_cr_read(cr, val);
-                        skip_emulated_instruction(vcpu);
-                        return 1;
+                        return kvm_skip_emulated_instruction(vcpu);
                 }
                 break;
         case 3: /* lmsw */
@@ -5732,8 +5737,7 @@ static int handle_cr(struct kvm_vcpu *vcpu)
                 trace_kvm_cr_write(0, (kvm_read_cr0(vcpu) & ~0xful) | val);
                 kvm_lmsw(vcpu, val);
 
-                skip_emulated_instruction(vcpu);
-                return 1;
+                return kvm_skip_emulated_instruction(vcpu);
         default:
                 break;
         }
@@ -5804,8 +5808,7 @@ static int handle_dr(struct kvm_vcpu *vcpu)
                 if (kvm_set_dr(vcpu, dr, kvm_register_readl(vcpu, reg)))
                         return 1;
 
-        skip_emulated_instruction(vcpu);
-        return 1;
+        return kvm_skip_emulated_instruction(vcpu);
 }
 
 static u64 vmx_get_dr6(struct kvm_vcpu *vcpu)
@@ -5858,8 +5861,7 @@ static int handle_rdmsr(struct kvm_vcpu *vcpu)
         /* FIXME: handling of bits 32:63 of rax, rdx */
         vcpu->arch.regs[VCPU_REGS_RAX] = msr_info.data & -1u;
         vcpu->arch.regs[VCPU_REGS_RDX] = (msr_info.data >> 32) & -1u;
-        skip_emulated_instruction(vcpu);
-        return 1;
+        return kvm_skip_emulated_instruction(vcpu);
 }
 
 static int handle_wrmsr(struct kvm_vcpu *vcpu)
@@ -5879,8 +5881,7 @@ static int handle_wrmsr(struct kvm_vcpu *vcpu)
         }
 
         trace_kvm_msr_write(ecx, data);
-        skip_emulated_instruction(vcpu);
-        return 1;
+        return kvm_skip_emulated_instruction(vcpu);
 }
 
 static int handle_tpr_below_threshold(struct kvm_vcpu *vcpu)
@@ -5924,8 +5925,7 @@ static int handle_invlpg(struct kvm_vcpu *vcpu)
         unsigned long exit_qualification = vmcs_readl(EXIT_QUALIFICATION);
 
         kvm_mmu_invlpg(vcpu, exit_qualification);
-        skip_emulated_instruction(vcpu);
-        return 1;
+        return kvm_skip_emulated_instruction(vcpu);
 }
 
 static int handle_rdpmc(struct kvm_vcpu *vcpu)
@@ -5933,15 +5933,12 @@ static int handle_rdpmc(struct kvm_vcpu *vcpu)
         int err;
 
         err = kvm_rdpmc(vcpu);
-        kvm_complete_insn_gp(vcpu, err);
-
-        return 1;
+        return kvm_complete_insn_gp(vcpu, err);
 }
 
 static int handle_wbinvd(struct kvm_vcpu *vcpu)
 {
-        kvm_emulate_wbinvd(vcpu);
-        return 1;
+        return kvm_emulate_wbinvd(vcpu);
 }
 
 static int handle_xsetbv(struct kvm_vcpu *vcpu)
@@ -5950,20 +5947,20 @@ static int handle_xsetbv(struct kvm_vcpu *vcpu)
         u32 index = kvm_register_read(vcpu, VCPU_REGS_RCX);
 
         if (kvm_set_xcr(vcpu, index, new_bv) == 0)
-                skip_emulated_instruction(vcpu);
+                return kvm_skip_emulated_instruction(vcpu);
         return 1;
 }
 
 static int handle_xsaves(struct kvm_vcpu *vcpu)
 {
-        skip_emulated_instruction(vcpu);
+        kvm_skip_emulated_instruction(vcpu);
         WARN(1, "this should never happen\n");
         return 1;
 }
 
 static int handle_xrstors(struct kvm_vcpu *vcpu)
 {
-        skip_emulated_instruction(vcpu);
+        kvm_skip_emulated_instruction(vcpu);
         WARN(1, "this should never happen\n");
         return 1;
 }
@@ -5984,8 +5981,7 @@ static int handle_apic_access(struct kvm_vcpu *vcpu)
                 if ((access_type == TYPE_LINEAR_APIC_INST_WRITE) &&
                     (offset == APIC_EOI)) {
                         kvm_lapic_set_eoi(vcpu);
-                        skip_emulated_instruction(vcpu);
-                        return 1;
+                        return kvm_skip_emulated_instruction(vcpu);
                 }
         }
         return emulate_instruction(vcpu, 0) == EMULATE_DONE;
@@ -6134,8 +6130,7 @@ static int handle_ept_misconfig(struct kvm_vcpu *vcpu)
         gpa = vmcs_read64(GUEST_PHYSICAL_ADDRESS);
         if (!kvm_io_bus_write(vcpu, KVM_FAST_MMIO_BUS, gpa, 0, NULL)) {
                 trace_kvm_fast_mmio(gpa);
-                skip_emulated_instruction(vcpu);
-                return 1;
+                return kvm_skip_emulated_instruction(vcpu);
         }
 
         ret = handle_mmio_page_fault(vcpu, gpa, true);
@@ -6508,15 +6503,12 @@ static int handle_pause(struct kvm_vcpu *vcpu)
                 grow_ple_window(vcpu);
 
         kvm_vcpu_on_spin(vcpu);
-        skip_emulated_instruction(vcpu);
-
-        return 1;
+        return kvm_skip_emulated_instruction(vcpu);
 }
 
 static int handle_nop(struct kvm_vcpu *vcpu)
 {
-        skip_emulated_instruction(vcpu);
-        return 1;
+        return kvm_skip_emulated_instruction(vcpu);
 }
 
 static int handle_mwait(struct kvm_vcpu *vcpu)
@@ -6823,8 +6815,7 @@ static int nested_vmx_check_vmptr(struct kvm_vcpu *vcpu, int exit_reason,
                  */
                 if (!PAGE_ALIGNED(vmptr) || (vmptr >> maxphyaddr)) {
                         nested_vmx_failInvalid(vcpu);
-                        skip_emulated_instruction(vcpu);
-                        return 1;
+                        return kvm_skip_emulated_instruction(vcpu);
                 }
 
                 page = nested_get_page(vcpu, vmptr);
@@ -6832,8 +6823,7 @@ static int nested_vmx_check_vmptr(struct kvm_vcpu *vcpu, int exit_reason,
                     *(u32 *)kmap(page) != VMCS12_REVISION) {
                         nested_vmx_failInvalid(vcpu);
                         kunmap(page);
-                        skip_emulated_instruction(vcpu);
-                        return 1;
+                        return kvm_skip_emulated_instruction(vcpu);
                 }
                 kunmap(page);
                 vmx->nested.vmxon_ptr = vmptr;
@@ -6842,30 +6832,26 @@ static int nested_vmx_check_vmptr(struct kvm_vcpu *vcpu, int exit_reason,
                 if (!PAGE_ALIGNED(vmptr) || (vmptr >> maxphyaddr)) {
                         nested_vmx_failValid(vcpu,
                                              VMXERR_VMCLEAR_INVALID_ADDRESS);
-                        skip_emulated_instruction(vcpu);
-                        return 1;
+                        return kvm_skip_emulated_instruction(vcpu);
                 }
 
                 if (vmptr == vmx->nested.vmxon_ptr) {
                         nested_vmx_failValid(vcpu,
                                              VMXERR_VMCLEAR_VMXON_POINTER);
-                        skip_emulated_instruction(vcpu);
-                        return 1;
+                        return kvm_skip_emulated_instruction(vcpu);
                 }
                 break;
         case EXIT_REASON_VMPTRLD:
                 if (!PAGE_ALIGNED(vmptr) || (vmptr >> maxphyaddr)) {
                         nested_vmx_failValid(vcpu,
                                              VMXERR_VMPTRLD_INVALID_ADDRESS);
-                        skip_emulated_instruction(vcpu);
-                        return 1;
+                        return kvm_skip_emulated_instruction(vcpu);
                 }
 
                 if (vmptr == vmx->nested.vmxon_ptr) {
                         nested_vmx_failValid(vcpu,
                                              VMXERR_VMCLEAR_VMXON_POINTER);
-                        skip_emulated_instruction(vcpu);
-                        return 1;
+                        return kvm_skip_emulated_instruction(vcpu);
                 }
                 break;
         default:
@@ -6921,8 +6907,7 @@ static int handle_vmon(struct kvm_vcpu *vcpu)
 
         if (vmx->nested.vmxon) {
                 nested_vmx_failValid(vcpu, VMXERR_VMXON_IN_VMX_ROOT_OPERATION);
-                skip_emulated_instruction(vcpu);
-                return 1;
+                return kvm_skip_emulated_instruction(vcpu);
         }
 
         if ((vmx->msr_ia32_feature_control & VMXON_NEEDED_FEATURES)
@@ -6963,8 +6948,7 @@ static int handle_vmon(struct kvm_vcpu *vcpu)
         vmx->nested.vmxon = true;
 
         nested_vmx_succeed(vcpu);
-        skip_emulated_instruction(vcpu);
-        return 1;
+        return kvm_skip_emulated_instruction(vcpu);
 
 out_shadow_vmcs:
         kfree(vmx->nested.cached_vmcs12);
@@ -7084,8 +7068,7 @@ static int handle_vmoff(struct kvm_vcpu *vcpu)
                 return 1;
         free_nested(to_vmx(vcpu));
         nested_vmx_succeed(vcpu);
-        skip_emulated_instruction(vcpu);
-        return 1;
+        return kvm_skip_emulated_instruction(vcpu);
 }
 
 /* Emulate the VMCLEAR instruction */
@@ -7125,8 +7108,7 @@ static int handle_vmclear(struct kvm_vcpu *vcpu)
         nested_free_vmcs02(vmx, vmptr);
 
         nested_vmx_succeed(vcpu);
-        skip_emulated_instruction(vcpu);
-        return 1;
+        return kvm_skip_emulated_instruction(vcpu);
 }
 
 static int nested_vmx_run(struct kvm_vcpu *vcpu, bool launch);
@@ -7340,18 +7322,15 @@ static int handle_vmread(struct kvm_vcpu *vcpu)
         if (!nested_vmx_check_permission(vcpu))
                 return 1;
 
-        if (!nested_vmx_check_vmcs12(vcpu)) {
-                skip_emulated_instruction(vcpu);
-                return 1;
-        }
+        if (!nested_vmx_check_vmcs12(vcpu))
+                return kvm_skip_emulated_instruction(vcpu);
 
         /* Decode instruction info and find the field to read */
         field = kvm_register_readl(vcpu, (((vmx_instruction_info) >> 28) & 0xf));
         /* Read the field, zero-extended to a u64 field_value */
         if (vmcs12_read_any(vcpu, field, &field_value) < 0) {
                 nested_vmx_failValid(vcpu, VMXERR_UNSUPPORTED_VMCS_COMPONENT);
-                skip_emulated_instruction(vcpu);
-                return 1;
+                return kvm_skip_emulated_instruction(vcpu);
         }
         /*
          * Now copy part of this value to register or memory, as requested.
@@ -7371,8 +7350,7 @@ static int handle_vmread(struct kvm_vcpu *vcpu)
         }
 
         nested_vmx_succeed(vcpu);
-        skip_emulated_instruction(vcpu);
-        return 1;
+        return kvm_skip_emulated_instruction(vcpu);
 }
 
 
@@ -7394,10 +7372,8 @@ static int handle_vmwrite(struct kvm_vcpu *vcpu)
         if (!nested_vmx_check_permission(vcpu))
                 return 1;
 
-        if (!nested_vmx_check_vmcs12(vcpu)) {
-                skip_emulated_instruction(vcpu);
-                return 1;
-        }
+        if (!nested_vmx_check_vmcs12(vcpu))
+                return kvm_skip_emulated_instruction(vcpu);
 
         if (vmx_instruction_info & (1u << 10))
                 field_value = kvm_register_readl(vcpu,
@@ -7418,19 +7394,16 @@ static int handle_vmwrite(struct kvm_vcpu *vcpu)
         if (vmcs_field_readonly(field)) {
                 nested_vmx_failValid(vcpu,
                         VMXERR_VMWRITE_READ_ONLY_VMCS_COMPONENT);
-                skip_emulated_instruction(vcpu);
-                return 1;
+                return kvm_skip_emulated_instruction(vcpu);
         }
 
         if (vmcs12_write_any(vcpu, field, field_value) < 0) {
                 nested_vmx_failValid(vcpu, VMXERR_UNSUPPORTED_VMCS_COMPONENT);
-                skip_emulated_instruction(vcpu);
-                return 1;
+                return kvm_skip_emulated_instruction(vcpu);
         }
 
         nested_vmx_succeed(vcpu);
-        skip_emulated_instruction(vcpu);
-        return 1;
+        return kvm_skip_emulated_instruction(vcpu);
 }
 
 /* Emulate the VMPTRLD instruction */
@@ -7451,8 +7424,7 @@ static int handle_vmptrld(struct kvm_vcpu *vcpu)
                 page = nested_get_page(vcpu, vmptr);
                 if (page == NULL) {
                         nested_vmx_failInvalid(vcpu);
-                        skip_emulated_instruction(vcpu);
-                        return 1;
+                        return kvm_skip_emulated_instruction(vcpu);
                 }
                 new_vmcs12 = kmap(page);
                 if (new_vmcs12->revision_id != VMCS12_REVISION) {
@@ -7460,8 +7432,7 @@ static int handle_vmptrld(struct kvm_vcpu *vcpu)
                         nested_release_page_clean(page);
                         nested_vmx_failValid(vcpu,
                                 VMXERR_VMPTRLD_INCORRECT_VMCS_REVISION_ID);
-                        skip_emulated_instruction(vcpu);
-                        return 1;
+                        return kvm_skip_emulated_instruction(vcpu);
                 }
 
                 nested_release_vmcs12(vmx);
@@ -7485,8 +7456,7 @@ static int handle_vmptrld(struct kvm_vcpu *vcpu)
         }
 
         nested_vmx_succeed(vcpu);
-        skip_emulated_instruction(vcpu);
-        return 1;
+        return kvm_skip_emulated_instruction(vcpu);
 }
 
 /* Emulate the VMPTRST instruction */
@@ -7511,8 +7481,7 @@ static int handle_vmptrst(struct kvm_vcpu *vcpu)
                 return 1;
         }
         nested_vmx_succeed(vcpu);
-        skip_emulated_instruction(vcpu);
-        return 1;
+        return kvm_skip_emulated_instruction(vcpu);
 }
 
 /* Emulate the INVEPT instruction */
@@ -7550,8 +7519,7 @@ static int handle_invept(struct kvm_vcpu *vcpu)
         if (type >= 32 || !(types & (1 << type))) {
                 nested_vmx_failValid(vcpu,
                                 VMXERR_INVALID_OPERAND_TO_INVEPT_INVVPID);
-                skip_emulated_instruction(vcpu);
-                return 1;
+                return kvm_skip_emulated_instruction(vcpu);
         }
 
         /* According to the Intel VMX instruction reference, the memory
@@ -7582,8 +7550,7 @@ static int handle_invept(struct kvm_vcpu *vcpu)
                 break;
         }
 
-        skip_emulated_instruction(vcpu);
-        return 1;
+        return kvm_skip_emulated_instruction(vcpu);
 }
 
 static int handle_invvpid(struct kvm_vcpu *vcpu)
@@ -7614,8 +7581,7 @@ static int handle_invvpid(struct kvm_vcpu *vcpu)
         if (type >= 32 || !(types & (1 << type))) {
                 nested_vmx_failValid(vcpu,
                         VMXERR_INVALID_OPERAND_TO_INVEPT_INVVPID);
-                skip_emulated_instruction(vcpu);
-                return 1;
+                return kvm_skip_emulated_instruction(vcpu);
         }
 
         /* according to the intel vmx instruction reference, the memory
@@ -7637,23 +7603,20 @@ static int handle_invvpid(struct kvm_vcpu *vcpu)
                 if (!vpid) {
                         nested_vmx_failValid(vcpu,
                                 VMXERR_INVALID_OPERAND_TO_INVEPT_INVVPID);
-                        skip_emulated_instruction(vcpu);
-                        return 1;
+                        return kvm_skip_emulated_instruction(vcpu);
                 }
                 break;
         case VMX_VPID_EXTENT_ALL_CONTEXT:
                 break;
         default:
                 WARN_ON_ONCE(1);
-                skip_emulated_instruction(vcpu);
-                return 1;
+                return kvm_skip_emulated_instruction(vcpu);
         }
 
         __vmx_flush_tlb(vcpu, vmx->nested.vpid02);
         nested_vmx_succeed(vcpu);
 
-        skip_emulated_instruction(vcpu);
-        return 1;
+        return kvm_skip_emulated_instruction(vcpu);
 }
 
 static int handle_pml_full(struct kvm_vcpu *vcpu)
@@ -10194,6 +10157,11 @@ static int nested_vmx_run(struct kvm_vcpu *vcpu, bool launch)
         if (!vmcs02)
                 return -ENOMEM;
 
+        /*
+         * After this point, the trap flag no longer triggers a singlestep trap
+         * on the vm entry instructions. Don't call
+         * kvm_skip_emulated_instruction.
+         */
         skip_emulated_instruction(vcpu);
         enter_guest_mode(vcpu);
 
@@ -10238,8 +10206,7 @@ static int nested_vmx_run(struct kvm_vcpu *vcpu, bool launch)
         return 1;
 
 out:
-        skip_emulated_instruction(vcpu);
-        return 1;
+        return kvm_skip_emulated_instruction(vcpu);
 }
 
 /*
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index ec59301f5192..7b38c5e6f412 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -425,12 +425,14 @@ void kvm_requeue_exception(struct kvm_vcpu *vcpu, unsigned nr)
 }
 EXPORT_SYMBOL_GPL(kvm_requeue_exception);
 
-void kvm_complete_insn_gp(struct kvm_vcpu *vcpu, int err)
+int kvm_complete_insn_gp(struct kvm_vcpu *vcpu, int err)
 {
         if (err)
                 kvm_inject_gp(vcpu, 0);
         else
-                kvm_x86_ops->skip_emulated_instruction(vcpu);
+                return kvm_skip_emulated_instruction(vcpu);
+
+        return 1;
 }
 EXPORT_SYMBOL_GPL(kvm_complete_insn_gp);
 
@@ -4813,8 +4815,8 @@ static int kvm_emulate_wbinvd_noskip(struct kvm_vcpu *vcpu)
 
 int kvm_emulate_wbinvd(struct kvm_vcpu *vcpu)
 {
-        kvm_x86_ops->skip_emulated_instruction(vcpu);
-        return kvm_emulate_wbinvd_noskip(vcpu);
+        kvm_emulate_wbinvd_noskip(vcpu);
+        return kvm_skip_emulated_instruction(vcpu);
 }
 EXPORT_SYMBOL_GPL(kvm_emulate_wbinvd);
 
@@ -5430,6 +5432,17 @@ static void kvm_vcpu_check_singlestep(struct kvm_vcpu *vcpu, unsigned long rflag
         }
 }
 
+int kvm_skip_emulated_instruction(struct kvm_vcpu *vcpu)
+{
+        unsigned long rflags = kvm_x86_ops->get_rflags(vcpu);
+        int r = EMULATE_DONE;
+
+        kvm_x86_ops->skip_emulated_instruction(vcpu);
+        kvm_vcpu_check_singlestep(vcpu, rflags, &r);
+        return r == EMULATE_DONE;
+}
+EXPORT_SYMBOL_GPL(kvm_skip_emulated_instruction);
+
 static bool kvm_vcpu_check_breakpoint(struct kvm_vcpu *vcpu, int *r)
 {
         if (unlikely(vcpu->guest_debug & KVM_GUESTDBG_USE_HW_BP) &&
@@ -6007,8 +6020,12 @@ EXPORT_SYMBOL_GPL(kvm_vcpu_halt);
 
 int kvm_emulate_halt(struct kvm_vcpu *vcpu)
 {
-        kvm_x86_ops->skip_emulated_instruction(vcpu);
-        return kvm_vcpu_halt(vcpu);
+        int ret = kvm_skip_emulated_instruction(vcpu);
+        /*
+         * TODO: we might be squashing a GUESTDBG_SINGLESTEP-triggered
+         * KVM_EXIT_DEBUG here.
+         */
+        return kvm_vcpu_halt(vcpu) && ret;
 }
 EXPORT_SYMBOL_GPL(kvm_emulate_halt);
 
@@ -6039,9 +6056,9 @@ void kvm_vcpu_deactivate_apicv(struct kvm_vcpu *vcpu)
 int kvm_emulate_hypercall(struct kvm_vcpu *vcpu)
 {
         unsigned long nr, a0, a1, a2, a3, ret;
-        int op_64_bit, r = 1;
+        int op_64_bit, r;
 
-        kvm_x86_ops->skip_emulated_instruction(vcpu);
+        r = kvm_skip_emulated_instruction(vcpu);
 
         if (kvm_hv_hypercall_enabled(vcpu->kvm))
                 return kvm_hv_hypercall(vcpu);