diff --git a/examples/host/msc_file_explorer/src/msc_app.c b/examples/host/msc_file_explorer/src/msc_app.c
index ddd39c674..035d74689 100644
--- a/examples/host/msc_file_explorer/src/msc_app.c
+++ b/examples/host/msc_file_explorer/src/msc_app.c
@@ -53,7 +53,11 @@ static CLI_UINT cli_buffer[BYTES_TO_CLI_UINTS(CLI_BUFFER_SIZE)];
 static FATFS fatfs[CFG_TUH_DEVICE_MAX]; // for simplicity only support 1 LUN per device
 static volatile bool _disk_busy[CFG_TUH_DEVICE_MAX];
 
-static scsi_inquiry_resp_t inquiry_resp;
+// define the buffer to be place in USB/DMA memory with correct alignment/cache line size
+CFG_TUH_MEM_SECTION static struct {
+  TUH_EPBUF_TYPE_DEF(scsi_inquiry_resp_t, inquiry);
+} scsi_resp;
+
 
 //--------------------------------------------------------------------+
 //
@@ -107,7 +111,7 @@ bool inquiry_complete_cb(uint8_t dev_addr, tuh_msc_complete_data_t const * cb_da
   }
 
   // Print out Vendor ID, Product ID and Rev
-  printf("%.8s %.16s rev %.4s\r\n", inquiry_resp.vendor_id, inquiry_resp.product_id, inquiry_resp.product_rev);
+  printf("%.8s %.16s rev %.4s\r\n", scsi_resp.inquiry.vendor_id, scsi_resp.inquiry.product_id, scsi_resp.inquiry.product_rev);
 
   // Get capacity of device
   uint32_t const block_count = tuh_msc_get_block_count(dev_addr, cbw->lun);
@@ -145,7 +149,7 @@ void tuh_msc_mount_cb(uint8_t dev_addr)
   printf("A MassStorage device is mounted\r\n");
 
   uint8_t const lun = 0;
-  tuh_msc_inquiry(dev_addr, lun, &inquiry_resp, inquiry_complete_cb, 0);
+  tuh_msc_inquiry(dev_addr, lun, &scsi_resp.inquiry, inquiry_complete_cb, 0);
 }
 
 void tuh_msc_umount_cb(uint8_t dev_addr)
diff --git a/src/class/msc/msc_host.c b/src/class/msc/msc_host.c
index 02791dace..900b31d7d 100644
--- a/src/class/msc/msc_host.c
+++ b/src/class/msc/msc_host.c
@@ -54,13 +54,12 @@ typedef struct {
   uint8_t itf_num;
   uint8_t ep_in;
   uint8_t ep_out;
-
   uint8_t max_lun;
 
   volatile bool configured; // Receive SET_CONFIGURE
   volatile bool mounted;    // Enumeration is complete
 
-  //------------- SCSI -------------//
+  // SCSI command data
   uint8_t stage;
   void* buffer;
   tuh_msc_complete_cb_t complete_cb;
@@ -82,7 +81,7 @@ CFG_TUH_MEM_SECTION static msch_epbuf_t _msch_epbuf[CFG_TUH_DEVICE_MAX];
 
 // Epbuf for enumeration, shared for all devices
 CFG_TUH_MEM_SECTION static struct {
-  TUH_EPBUF_DEF(buf, 32); // TODO make this configurable
+  TUH_EPBUF_DEF(buf, 32);
 } _msch_enum_buf;
 
 TU_ATTR_ALWAYS_INLINE static inline msch_interface_t* get_itf(uint8_t daddr) {
@@ -141,10 +140,10 @@ bool tuh_msc_scsi_command(uint8_t daddr, msc_cbw_t const* cbw, void* data,
   msch_epbuf_t* epbuf = get_epbuf(daddr);
 
   epbuf->cbw = *cbw;
-  p_msc->stage = MSC_STAGE_CMD;
   p_msc->buffer = data;
   p_msc->complete_cb = complete_cb;
   p_msc->complete_arg = arg;
+  p_msc->stage = MSC_STAGE_CMD;
 
   if (!usbh_edpt_xfer(daddr, p_msc->ep_out, (uint8_t*) &epbuf->cbw, sizeof(msc_cbw_t))) {
     usbh_edpt_release(daddr, p_msc->ep_out);
@@ -292,6 +291,7 @@ bool tuh_msc_reset(uint8_t dev_addr) {
 //--------------------------------------------------------------------+
 bool msch_init(void) {
   TU_LOG_DRV("sizeof(msch_interface_t) = %u\r\n", sizeof(msch_interface_t));
+  TU_LOG_DRV("sizeof(msch_epbuf_t) = %u\r\n", sizeof(msch_epbuf_t));
   tu_memclr(_msch_itf, sizeof(_msch_itf));
   return true;
 }
@@ -325,18 +325,15 @@ bool msch_xfer_cb(uint8_t dev_addr, uint8_t ep_addr, xfer_result_t event, uint32
     case MSC_STAGE_CMD:
       // Must be Command Block
       TU_ASSERT(ep_addr == p_msc->ep_out && event == XFER_RESULT_SUCCESS && xferred_bytes == sizeof(msc_cbw_t));
-
       if (cbw->total_bytes && p_msc->buffer) {
         // Data stage if any
         p_msc->stage = MSC_STAGE_DATA;
         uint8_t const ep_data = (cbw->dir & TUSB_DIR_IN_MASK) ? p_msc->ep_in : p_msc->ep_out;
         TU_ASSERT(usbh_edpt_xfer(dev_addr, ep_data, p_msc->buffer, (uint16_t) cbw->total_bytes));
-      } else {
-        // Status stage
-        p_msc->stage = MSC_STAGE_STATUS;
-        TU_ASSERT(usbh_edpt_xfer(dev_addr, p_msc->ep_in, (uint8_t*) csw, (uint16_t) sizeof(msc_csw_t)));
+        break;
       }
-      break;
+
+      TU_ATTR_FALLTHROUGH; // fallthrough to status stage
 
     case MSC_STAGE_DATA:
       // Status stage
diff --git a/src/class/msc/msc_host.h b/src/class/msc/msc_host.h
index 6969336a0..09d777066 100644
--- a/src/class/msc/msc_host.h
+++ b/src/class/msc/msc_host.h
@@ -73,10 +73,12 @@ uint32_t tuh_msc_get_block_size(uint8_t dev_addr, uint8_t lun);
 // Perform a full SCSI command (cbw, data, csw) in non-blocking manner.
 // Complete callback is invoked when SCSI op is complete.
 // return true if success, false if there is already pending operation.
+// NOTE: buffer must be accessible by USB/DMA controller, aligned correctly and multiple of cache line if enabled
 bool tuh_msc_scsi_command(uint8_t daddr, msc_cbw_t const* cbw, void* data, tuh_msc_complete_cb_t complete_cb, uintptr_t arg);
 
 // Perform SCSI Inquiry command
 // Complete callback is invoked when SCSI op is complete.
+// NOTE: response must be accessible by USB/DMA controller, aligned correctly and multiple of cache line if enabled
 bool tuh_msc_inquiry(uint8_t dev_addr, uint8_t lun, scsi_inquiry_resp_t* response, tuh_msc_complete_cb_t complete_cb, uintptr_t arg);
 
 // Perform SCSI Test Unit Ready command
@@ -85,14 +87,17 @@ bool tuh_msc_test_unit_ready(uint8_t dev_addr, uint8_t lun, tuh_msc_complete_cb_
 
 // Perform SCSI Request Sense 10 command
 // Complete callback is invoked when SCSI op is complete.
+// NOTE: response must be accessible by USB/DMA controller, aligned correctly and multiple of cache line if enabled
 bool tuh_msc_request_sense(uint8_t dev_addr, uint8_t lun, void *response, tuh_msc_complete_cb_t complete_cb, uintptr_t arg);
 
 // Perform SCSI Read 10 command. Read n blocks starting from LBA to buffer
 // Complete callback is invoked when SCSI op is complete.
+// NOTE: buffer must be accessible by USB/DMA controller, aligned correctly and multiple of cache line if enabled
 bool tuh_msc_read10(uint8_t dev_addr, uint8_t lun, void * buffer, uint32_t lba, uint16_t block_count, tuh_msc_complete_cb_t complete_cb, uintptr_t arg);
 
 // Perform SCSI Write 10 command. Write n blocks starting from LBA to device
 // Complete callback is invoked when SCSI op is complete.
+// NOTE: buffer must be accessible by USB/DMA controller, aligned correctly and multiple of cache line if enabled
 bool tuh_msc_write10(uint8_t dev_addr, uint8_t lun, void const * buffer, uint32_t lba, uint16_t block_count, tuh_msc_complete_cb_t complete_cb, uintptr_t arg);
 
 // Perform SCSI Read Capacity 10 command