diff --git a/hypervisor/arch/x86/guest/vmcall.c b/hypervisor/arch/x86/guest/vmcall.c
index 4e80a37ed..d3deba1af 100644
--- a/hypervisor/arch/x86/guest/vmcall.c
+++ b/hypervisor/arch/x86/guest/vmcall.c
@@ -103,7 +103,7 @@ int vmcall_vmexit_handler(struct vcpu *vcpu)
 	case HC_NOTIFY_REQUEST_FINISH:
 		/* param1: vmid
 		 * param2: vcpu_id */
-		ret = hcall_notify_req_finish((uint16_t)param1,
+		ret = hcall_notify_ioreq_finish((uint16_t)param1,
 			(uint16_t)param2);
 		break;
 
diff --git a/hypervisor/common/hypercall.c b/hypervisor/common/hypercall.c
index e2ca3bf72..6d38187a4 100644
--- a/hypervisor/common/hypercall.c
+++ b/hypervisor/common/hypercall.c
@@ -361,20 +361,27 @@ int32_t hcall_set_ioreq_buffer(struct vm *vm, uint16_t vmid, uint64_t param)
 	return ret;
 }
 
-static void complete_request(struct vcpu *vcpu)
+static void complete_ioreq(struct vcpu *vcpu)
+{
+	union vhm_request_buffer *req_buf;
+	struct vhm_request *vhm_req;
+
+	req_buf = (union vhm_request_buffer *)
+		vcpu->vm->sw.io_shared_page;
+	vhm_req = &req_buf->req_queue[vcpu->vcpu_id];
+
+	vhm_req->valid = 0;
+	atomic_store32(&vcpu->ioreq_pending, 0U);
+}
+
+static void emulate_io_post(struct vcpu *vcpu)
 {
 	/*
 	 * If vcpu is in Zombie state and will be destroyed soon. Just
 	 * mark ioreq done and don't resume vcpu.
 	 */
 	if (vcpu->state == VCPU_ZOMBIE) {
-		union vhm_request_buffer *req_buf;
-
-		req_buf = (union vhm_request_buffer *)
-				vcpu->vm->sw.io_shared_page;
-		req_buf->req_queue[vcpu->vcpu_id].valid = false;
-		atomic_store32(&vcpu->ioreq_pending, 0U);
-
+		complete_ioreq(vcpu);
 		return;
 	}
 
@@ -384,17 +391,27 @@ static void complete_request(struct vcpu *vcpu)
 		break;
 
 	case REQ_PORTIO:
+	case REQ_PCICFG:
+		/* REQ_PORTIO on 0xcf8 & 0xcfc may switch to REQ_PCICFG in some
+		 * cases. It works to apply the post-work for REQ_PORTIO on
+		 * REQ_PCICFG because the format of the first 28 bytes of
+		 * REQ_PORTIO & REQ_PCICFG requests are exactly the same and
+		 * post-work is mainly interested in the read value.
+		 */
 		dm_emulate_pio_post(vcpu);
 		break;
 
 	default:
+		/* REQ_WP can only be triggered on writes which do not need
+		 * post-work. Just mark the ioreq done. */
+		complete_ioreq(vcpu);
 		break;
 	}
 
 	resume_vcpu(vcpu);
 }
 
-int32_t hcall_notify_req_finish(uint16_t vmid, uint16_t vcpu_id)
+int32_t hcall_notify_ioreq_finish(uint16_t vmid, uint16_t vcpu_id)
 {
 	union vhm_request_buffer *req_buf;
 	struct vhm_request *req;
@@ -423,7 +440,7 @@ int32_t hcall_notify_req_finish(uint16_t vmid, uint16_t vcpu_id)
 	if ((req->valid != 0) &&
 		((req->processed == REQ_STATE_SUCCESS) ||
 		 (req->processed == REQ_STATE_FAILED))) {
-		complete_request(vcpu);
+		emulate_io_post(vcpu);
 	}
 
 	return 0;
diff --git a/hypervisor/include/common/hypercall.h b/hypervisor/include/common/hypercall.h
index 3a5e9a512..f7b5ed83d 100644
--- a/hypervisor/include/common/hypercall.h
+++ b/hypervisor/include/common/hypercall.h
@@ -203,7 +203,7 @@ int32_t hcall_set_ioreq_buffer(struct vm *vm, uint16_t vmid, uint64_t param);
  *
  * @return 0 on success, non-zero on error.
  */
-int32_t hcall_notify_req_finish(uint16_t vmid, uint16_t vcpu_id);
+int32_t hcall_notify_ioreq_finish(uint16_t vmid, uint16_t vcpu_id);
 
 /**
  * @brief setup ept memory mapping