diff --git a/hypervisor/arch/x86/io.c b/hypervisor/arch/x86/io.c
index a7ca57115..660640c25 100644
--- a/hypervisor/arch/x86/io.c
+++ b/hypervisor/arch/x86/io.c
@@ -10,7 +10,6 @@
 
 static void complete_ioreq(struct vhm_request *vhm_req)
 {
-	vhm_req->valid = 0;
 	atomic_store32(&vhm_req->processed, REQ_STATE_FREE);
 }
 
@@ -140,8 +139,7 @@ void emulate_io_post(struct acrn_vcpu *vcpu)
 	req_buf = (union vhm_request_buffer *)vcpu->vm->sw.io_shared_page;
 	vhm_req = &req_buf->req_queue[vcpu->vcpu_id];
 
-	if ((vhm_req->valid == 0) ||
-		(atomic_load32(&vhm_req->processed) != REQ_STATE_COMPLETE)) {
+	if (atomic_load32(&vhm_req->processed) != REQ_STATE_COMPLETE) {
 		return;
 	}
 
diff --git a/hypervisor/common/io_request.c b/hypervisor/common/io_request.c
index 92aaaf386..c94a75d92 100644
--- a/hypervisor/common/io_request.c
+++ b/hypervisor/common/io_request.c
@@ -69,7 +69,6 @@ void reset_vm_ioreqs(struct acrn_vm *vm)
 
 	req_buf = vm->sw.io_shared_page;
 	for (i = 0U; i < VHM_REQUEST_MAX; i++) {
-		req_buf->req_queue[i].valid = 0U;
 		atomic_store32(&req_buf->req_queue[i].processed, REQ_STATE_FREE);
 	}
 }
@@ -84,11 +83,8 @@ static bool has_complete_ioreq(struct acrn_vcpu *vcpu)
 	req_buf = (union vhm_request_buffer *)vm->sw.io_shared_page;
 	if (req_buf != NULL) {
 		vhm_req = &req_buf->req_queue[vcpu->vcpu_id];
-		if (vhm_req->valid &&
-			atomic_load32(&vhm_req->processed)
-				== REQ_STATE_COMPLETE) {
+		if (atomic_load32(&vhm_req->processed) == REQ_STATE_COMPLETE) {
 			return true;
-
 		}
 	}
 
@@ -153,7 +149,7 @@ int32_t acrn_insert_request_wait(struct acrn_vcpu *vcpu, const struct io_request
 
 	/* pause vcpu, wait for VHM to handle the MMIO request.
 	 * TODO: when pause_vcpu changed to switch vcpu out directlly, we
-	 * should fix the race issue between req.valid = true and vcpu pause
+	 * should fix the race issue between req.processed update and vcpu pause
 	 */
 	pause_vcpu(vcpu, VCPU_PAUSED);
 
@@ -162,7 +158,6 @@ int32_t acrn_insert_request_wait(struct acrn_vcpu *vcpu, const struct io_request
 	 * before we perform upcall.
 	 * because VHM can work in pulling mode without wait for upcall
 	 */
-	vhm_req->valid = 1;
 	atomic_store32(&vhm_req->processed, REQ_STATE_PENDING);
 
 	acrn_print_request(vcpu->vcpu_id, vhm_req);
diff --git a/hypervisor/include/public/acrn_common.h b/hypervisor/include/public/acrn_common.h
index 073a155df..b280a29c7 100644
--- a/hypervisor/include/public/acrn_common.h
+++ b/hypervisor/include/public/acrn_common.h
@@ -261,7 +261,7 @@ union vhm_io_request {
  */
 struct vhm_request {
 	/**
-	 * Type of this request.
+	 * @brief Type of this request.
 	 *
 	 * Byte offset: 0.
 	 */
@@ -275,14 +275,16 @@ struct vhm_request {
 	uint32_t completion_polling;
 
 	/**
-	 * Reserved.
+	 * @brief Reserved.
 	 *
 	 * Byte offset: 8.
 	 */
 	uint32_t reserved0[14];
 
 	/**
-	 * Details about this request. For REQ_PORTIO, this has type
+	 * @brief Details about this request.
+	 *
+	 * For REQ_PORTIO, this has type
 	 * pio_request. For REQ_MMIO and REQ_WP, this has type mmio_request. For
 	 * REQ_PCICFG, this has type pci_request.
 	 *
@@ -291,25 +293,25 @@ struct vhm_request {
 	union vhm_io_request reqs;
 
 	/**
-	 * Whether this request is valid for processing. ACRN write, VHM read
-	 * only.
-	 *
-	 * Warning; this field is obsolete and will be removed soon.
+	 * @brief Reserved.
 	 *
 	 * Byte offset: 128.
 	 */
-	int32_t valid;
+	uint32_t reserved1;
 
 	/**
-	 * The client which is distributed to handle this request. Accessed by
-	 * VHM only.
+	 * @brief The client which is distributed to handle this request.
+	 *
+	 * Accessed by VHM only.
 	 *
 	 * Byte offset: 132.
 	 */
 	int32_t client;
 
 	/**
-	 * The status of this request, taking REQ_STATE_xxx as values.
+	 * @brief The status of this request.
+	 *
+	 * Taking REQ_STATE_xxx as values.
 	 *
 	 * Byte offset: 136.
 	 */