diff --git a/core/app-mgr/app-manager/module_wasm_app.c b/core/app-mgr/app-manager/module_wasm_app.c
index 72e1f5e6f..a3e7a2a76 100644
--- a/core/app-mgr/app-manager/module_wasm_app.c
+++ b/core/app-mgr/app-manager/module_wasm_app.c
@@ -13,7 +13,9 @@
 #include "event.h"
 #include "watchdog.h"
 #include "runtime_lib.h"
+#if WASM_ENABLE_INTERP != 0
 #include "wasm.h"
+#endif
 #if WASM_ENABLE_AOT != 0
 #include "aot_export.h"
 #endif
@@ -163,6 +165,15 @@ module_interface wasm_app_module_interface = {
     wasm_app_module_on_install_request_byte_arrive
 };
 
+#if WASM_ENABLE_INTERP == 0
+static unsigned
+align_uint(unsigned v, unsigned b)
+{
+    unsigned m = b - 1;
+    return (v + m) & ~m;
+}
+#endif
+
 static void
 exchange_uint32(uint8 *p_data)
 {
@@ -577,7 +588,7 @@ wasm_app_module_install(request_t * msg)
     char m_name[APP_NAME_MAX_LEN] = { 0 };
     char timeout_str[MAX_INT_STR_LEN] = { 0 };
     char heap_size_str[MAX_INT_STR_LEN] = { 0 };
-    char timers_str[MAX_INT_STR_LEN] = { 0 }, err[256];
+    char timers_str[MAX_INT_STR_LEN] = { 0 }, err[128], err_resp[256];
 #if WASM_ENABLE_LIBC_WASI != 0
     char wasi_dir_buf[PATH_MAX] = { 0 };
     const char *wasi_dir_list[] = { wasi_dir_buf };
@@ -651,8 +662,9 @@ wasm_app_module_install(request_t * msg)
             module = wasm_runtime_load_from_sections(aot_file->sections, true,
                                                      err, err_size);
             if (!module) {
-                SEND_ERR_RESPONSE(msg->mid,
-                                  "Install WASM app failed: load WASM file failed.");
+                snprintf(err_resp, sizeof(err_resp),
+                         "Install WASM app failed: %s", err);
+                SEND_ERR_RESPONSE(msg->mid, err_resp);
                 goto fail;
             }
             /* Destroy useless sections from list after load */
@@ -677,8 +689,9 @@ wasm_app_module_install(request_t * msg)
             /* Instantiate the AOT module */
             inst = wasm_runtime_instantiate(module, 0, heap_size, err, err_size);
             if (!inst) {
-                SEND_ERR_RESPONSE(msg->mid,
-                                  "Install WASM app failed: instantiate wasm runtime failed.");
+                snprintf(err_resp, sizeof(err_resp),
+                         "Install WASM app failed: %s", err);
+                SEND_ERR_RESPONSE(msg->mid, err);
                 goto fail;
             }
             break;
@@ -715,8 +728,9 @@ wasm_app_module_install(request_t * msg)
             module = wasm_runtime_load_from_sections(bytecode_file->sections, false,
                                                      err, err_size);
             if (!module) {
-                SEND_ERR_RESPONSE(msg->mid,
-                                  "Install WASM app failed: load WASM file failed.");
+                snprintf(err_resp, sizeof(err_resp),
+                         "Install WASM app failed: %s", err);
+                SEND_ERR_RESPONSE(msg->mid, err_resp);
                 goto fail;
             }
 
@@ -742,8 +756,9 @@ wasm_app_module_install(request_t * msg)
             /* Instantiate the wasm module */
             inst = wasm_runtime_instantiate(module, 0, heap_size, err, err_size);
             if (!inst) {
-                SEND_ERR_RESPONSE(msg->mid,
-                                  "Install WASM app failed: instantiate wasm runtime failed.");
+                snprintf(err_resp, sizeof(err_resp),
+                         "Install WASM app failed: %s", err);
+                SEND_ERR_RESPONSE(msg->mid, err_resp);
                 goto fail;
             }
 
diff --git a/core/shared/platform/zephyr/zephyr_thread.c b/core/shared/platform/zephyr/zephyr_thread.c
index 932255195..55cadb47c 100644
--- a/core/shared/platform/zephyr/zephyr_thread.c
+++ b/core/shared/platform/zephyr/zephyr_thread.c
@@ -14,6 +14,55 @@
     }                                                       \
   } while (0)
 
+#if defined(CONFIG_ARM_MPU) || defined(CONFIG_ARC_MPU) \
+    || KERNEL_VERSION_NUMBER > 0x020300 /* version 2.3.0 */
+#define BH_ENABLE_ZEPHYR_MPU_STACK 1
+#elif !defined(BH_ENABLE_ZEPHYR_MPU_STACK)
+#define BH_ENABLE_ZEPHYR_MPU_STACK 0
+#endif
+#if !defined(BH_ZEPHYR_MPU_STACK_SIZE)
+#define BH_ZEPHYR_MPU_STACK_SIZE APP_THREAD_STACK_SIZE_MIN
+#endif
+#if !defined(BH_ZEPHYR_MPU_STACK_COUNT)
+#define BH_ZEPHYR_MPU_STACK_COUNT 4
+#endif
+
+#if BH_ENABLE_ZEPHYR_MPU_STACK != 0
+static K_THREAD_STACK_ARRAY_DEFINE(mpu_stacks,
+                                   BH_ZEPHYR_MPU_STACK_COUNT,
+                                   BH_ZEPHYR_MPU_STACK_SIZE);
+static bool mpu_stack_allocated[BH_ZEPHYR_MPU_STACK_COUNT];
+static struct k_mutex mpu_stack_lock;
+
+static char *mpu_stack_alloc()
+{
+    int i;
+
+    k_mutex_lock(&mpu_stack_lock, K_FOREVER);
+    for (i = 0; i < BH_ZEPHYR_MPU_STACK_COUNT; i++) {
+        if (!mpu_stack_allocated[i]) {
+            mpu_stack_allocated[i] = true;
+            k_mutex_unlock(&mpu_stack_lock);
+            return (char*)mpu_stacks[i];
+        }
+    }
+    k_mutex_unlock(&mpu_stack_lock);
+    return NULL;
+}
+
+static void mpu_stack_free(char *stack)
+{
+    int i;
+
+    k_mutex_lock(&mpu_stack_lock, K_FOREVER);
+    for (i = 0; i < BH_ZEPHYR_MPU_STACK_COUNT; i++) {
+        if ((char *)mpu_stacks[i] == stack)
+            mpu_stack_allocated[i] = false;
+    }
+    k_mutex_unlock(&mpu_stack_lock);
+}
+#endif
+
 typedef struct os_thread_wait_node {
     struct k_sem sem;
     os_thread_wait_list next;
@@ -32,8 +81,12 @@ typedef struct os_thread_data {
     os_thread_wait_list thread_wait_list;
     /* Thread stack size */
     unsigned stack_size;
+#if BH_ENABLE_ZEPHYR_MPU_STACK == 0
     /* Thread stack */
     char stack[1];
+#else
+    char *stack;
+#endif
 } os_thread_data;
 
 typedef struct os_thread_obj {
@@ -164,6 +217,9 @@ int os_thread_sys_init()
     if (is_thread_sys_inited)
         return BHT_OK;
 
+#if BH_ENABLE_ZEPHYR_MPU_STACK != 0
+    k_mutex_init(&mpu_stack_lock);
+#endif
     k_mutex_init(&thread_data_lock);
     k_mutex_init(&thread_obj_lock);
 
@@ -214,6 +270,9 @@ static void os_thread_cleanup(void)
     /* Set flag to true for the next thread creating to
      free the thread object */
     ((os_thread_obj*) thread_data->tid)->to_be_freed = true;
+#if BH_ENABLE_ZEPHYR_MPU_STACK != 0
+    mpu_stack_free(thread_data->stack);
+#endif
     BH_FREE(thread_data);
 }
 
@@ -253,14 +312,17 @@ int os_thread_create_with_prio(korp_tid *p_tid, thread_start_routine_t start,
 
     memset(tid, 0, sizeof(os_thread_obj));
 
+    /* Create and initialize thread data */
+#if BH_ENABLE_ZEPHYR_MPU_STACK == 0
     if (stack_size < APP_THREAD_STACK_SIZE_MIN)
         stack_size = APP_THREAD_STACK_SIZE_MIN;
-
-    /* Create and initialize thread data */
     thread_data_size = offsetof(os_thread_data, stack) + stack_size;
+#else
+    stack_size = BH_ZEPHYR_MPU_STACK_SIZE;
+    thread_data_size = sizeof(os_thread_data);
+#endif
     if (!(thread_data = BH_MALLOC(thread_data_size))) {
-        BH_FREE(tid);
-        return BHT_ERROR;
+        goto fail1;
     }
 
     memset(thread_data, 0, thread_data_size);
@@ -268,22 +330,36 @@ int os_thread_create_with_prio(korp_tid *p_tid, thread_start_routine_t start,
     thread_data->stack_size = stack_size;
     thread_data->tid = tid;
 
+#if BH_ENABLE_ZEPHYR_MPU_STACK != 0
+    if (!(thread_data->stack = mpu_stack_alloc())) {
+        goto fail2;
+    }
+#endif
+
     /* Create the thread */
     if (!((tid = k_thread_create(tid, (k_thread_stack_t *)thread_data->stack,
                                  stack_size, os_thread_wrapper, start, arg,
                                  thread_data, prio, 0, K_NO_WAIT)))) {
-        BH_FREE(tid);
-        BH_FREE(thread_data);
-        return BHT_ERROR;
+        goto fail3;
     }
 
     bh_assert(tid == thread_data->tid);
 
     /* Set thread custom data */
     thread_data_list_add(thread_data);
-    thread_obj_list_add((os_thread_obj*) tid);
+    thread_obj_list_add((os_thread_obj*)tid);
     *p_tid = tid;
     return BHT_OK;
+
+fail3:
+#if BH_ENABLE_ZEPHYR_MPU_STACK != 0
+    mpu_stack_free(thread_data->stack);
+fail2:
+#endif
+    BH_FREE(thread_data);
+fail1:
+    BH_FREE(tid);
+    return BHT_ERROR;
 }
 
 korp_tid os_self_thread()
diff --git a/samples/gui/README.md b/samples/gui/README.md
index 2d3d062cb..d79453c3e 100644
--- a/samples/gui/README.md
+++ b/samples/gui/README.md
@@ -77,7 +77,7 @@ Refer to  [Zephyr getting started](https://docs.zephyrproject.org/latest/getting
  b. copy samples
   ```bash
   cd zephyr/samples
-  cp -a <wamr_root>samples/gui/wasm-runtime-wgl wasm-runtime-wgl
+  cp -a <wamr_root>/samples/gui/wasm-runtime-wgl wasm-runtime-wgl
   cd wasm-runtime-wgl/zephyr_build
   ```
  c. create a link to wamr root dir
@@ -123,7 +123,8 @@ Refer to  [Zephyr getting started](https://docs.zephyrproject.org/latest/getting
 - Install WASM application to Zephyr using host_tool
 First, connect PC and STM32 with UART. Then install to use host_tool.
   ```bash
-  ./host_tool -D /dev/ttyUSBXXX -i inc -f ui_increase.wasm
+  sudo ./host_tool -D /dev/ttyUSBXXX -i inc -f ui_increase.wasm
+  # /dev/ttyUSBXXX is the UART device, e.g. /dev/ttyUSB0
   ```
 
 - Install AOT version WASM application
diff --git a/samples/littlevgl/README.md b/samples/littlevgl/README.md
index 0b6bbe9fb..87fd2bd42 100644
--- a/samples/littlevgl/README.md
+++ b/samples/littlevgl/README.md
@@ -103,7 +103,7 @@ We can use a STM32 NUCLEO_F767ZI  board with ILI9341 display and XPT2046 touch s
   b. copy samples
   ```bash
   cd zephyr/samples/
-  cp -a <wamr_root>samples/littlevgl/vgl-wasm-runtime vgl-wasm-runtime
+  cp -a <wamr_root>/samples/littlevgl/vgl-wasm-runtime vgl-wasm-runtime
   cd vgl-wasm-runtime/zephyr_build
   ```
   c. create a link to wamr root dir
@@ -161,7 +161,8 @@ d. build source code
 - Install WASM application to Zephyr using host_tool
 First, connect PC and STM32 with UART. Then install to use host_tool.
   ```bash
-  ./host_tool -D /dev/ttyUSBXXX -i ui_app -f ui_app_builtin_libc.wasm
+  sudo ./host_tool -D /dev/ttyUSBXXX -i ui_app -f ui_app_builtin_libc.wasm
+  # /dev/ttyUSBXXX is the UART device, e.g. /dev/ttyUSB0
   ```
 **Note**: WASI is unavailable on zephyr currently, so you have to use the ui_app_builtin_libc.wasm which doesn't depend on WASI.
 
diff --git a/samples/littlevgl/vgl-wasm-runtime/src/platform/zephyr/display_ili9340.c b/samples/littlevgl/vgl-wasm-runtime/src/platform/zephyr/display_ili9340.c
index eb4295094..4a0276787 100644
--- a/samples/littlevgl/vgl-wasm-runtime/src/platform/zephyr/display_ili9340.c
+++ b/samples/littlevgl/vgl-wasm-runtime/src/platform/zephyr/display_ili9340.c
@@ -26,7 +26,7 @@ struct ili9340_data {
     struct device *spi_dev;
     struct spi_config spi_config;
 #ifdef DT_ILITEK_ILI9340_0_CS_GPIO_CONTROLLER
-struct spi_cs_control cs_ctrl;
+    struct spi_cs_control cs_ctrl;
 #endif
 };
 
diff --git a/samples/littlevgl/vgl-wasm-runtime/src/platform/zephyr/iwasm_main.c b/samples/littlevgl/vgl-wasm-runtime/src/platform/zephyr/iwasm_main.c
index e2f4a5c9e..6ec3bd49e 100644
--- a/samples/littlevgl/vgl-wasm-runtime/src/platform/zephyr/iwasm_main.c
+++ b/samples/littlevgl/vgl-wasm-runtime/src/platform/zephyr/iwasm_main.c
@@ -76,7 +76,7 @@ host_interface interface = {
 
 timer_ctx_t timer_ctx;
 
-static char global_heap_buf[368 * 1024] = { 0 };
+static char global_heap_buf[350 * 1024] = { 0 };
 
 static NativeSymbol native_symbols[] = {
     EXPORT_WASM_API_WITH_SIG(display_input_read, "(*)i"),
diff --git a/samples/littlevgl/vgl-wasm-runtime/zephyr-build/prj.conf b/samples/littlevgl/vgl-wasm-runtime/zephyr-build/prj.conf
index f9f13f9a7..6ca7f4426 100644
--- a/samples/littlevgl/vgl-wasm-runtime/zephyr-build/prj.conf
+++ b/samples/littlevgl/vgl-wasm-runtime/zephyr-build/prj.conf
@@ -1,6 +1,5 @@
 CONFIG_SPI=y
 CONFIG_SPI_STM32=y
-CONFIG_SPI_1=y
 CONFIG_PRINTK=y
 CONFIG_LOG=y
 #CONFIG_UART_2=y
diff --git a/test-tools/host-tool/src/main.c b/test-tools/host-tool/src/main.c
index 230977b4c..305abc7ad 100644
--- a/test-tools/host-tool/src/main.c
+++ b/test-tools/host-tool/src/main.c
@@ -750,7 +750,7 @@ static void output_event(request_t *obj)
 
 int main(int argc, char *argv[])
 {
-    int ret;
+    int ret = -1;
     imrt_link_recv_context_t recv_ctx = { 0 };
     char buffer[BUF_SIZE] = { 0 };
     uint32_t last_check = 0, total_elpased_ms = 0;