diff --git a/cortex_handlers.c b/cortex_handlers.c
index cbf63b0..8599f03 100644
--- a/cortex_handlers.c
+++ b/cortex_handlers.c
@@ -37,7 +37,7 @@ void vApplicationStackOverflowHook(TaskHandle_t xTask, signed char *pcTaskName)
 
 // && (__CORTEX_M >= 3)
 #if VERBOSE_HARDFAULT
-void prvGetRegistersFromStack( uint32_t *origStack, uint32_t lr_value)
+void __attribute__((used)) HardFault_DumpRegisters( uint32_t *origStack, uint32_t lr_value)
 {
 /* These are volatile to try and prevent the compiler/linker optimising them
 away as the variables never actually get used.  If the debugger won't show the
@@ -179,7 +179,7 @@ void  __attribute__((naked)) HardFault_Handler(void)
 //        " mov r2, lr                                                \n"
 //        " ldr r3, handler2_address_const                            \n"
 //        " bx r3                                                     \n"
-//        " handler2_address_const: .word prvGetRegistersFromStack    \n"
+//        " handler2_address_const: .word HardFault_DumpRegisters     \n"
 //    );
 //
     __asm volatile(  ".syntax unified\n"
@@ -188,10 +188,10 @@ void  __attribute__((naked)) HardFault_Handler(void)
         "TST    R0, R1  \n"
         "BEQ    _MSP    \n"
         "MRS    R0, PSP \n"
-        "B      prvGetRegistersFromStack      \n"
+        "B      HardFault_DumpRegisters       \n"
         "_MSP:  \n"
         "MRS    R0, MSP \n"
-        "B      prvGetRegistersFromStack      \n"
+        "B      HardFault_DumpRegisters       \n"
         ".syntax divided\n") ;
 
 #endif
diff --git a/freertos.c b/freertos.c
index 1de37d0..1b2e0a4 100644
--- a/freertos.c
+++ b/freertos.c
@@ -59,15 +59,15 @@
 /* Variables -----------------------------------------------------------------*/
 
 osThreadId tskMainHandle;
-uint32_t mainTaskBuffer[ TSK_STACK_MAIN ];
+uint32_t mainTaskStack[ TSK_STACK_MAIN ];
 osStaticThreadDef_t mainTaskControlBlock;
 
 osThreadId tskMsgHandle;
-uint32_t msgTaskBuffer[ TSK_STACK_MSG ];
+uint32_t msgTaskStack[ TSK_STACK_MSG ];
 osStaticThreadDef_t msgTaskControlBlock;
 
 osThreadId tskJobRunnerHandle;
-uint32_t jobRunnerBuffer[ TSK_STACK_JOBRUNNER ];
+uint32_t jobRunnerStack[ TSK_STACK_JOBRUNNER ];
 osStaticThreadDef_t jobRunnerControlBlock;
 
 osMessageQId queSchedHandle;
@@ -135,9 +135,9 @@ void vApplicationGetIdleTaskMemory( StaticTask_t **ppxIdleTaskTCBBuffer, StackTy
 
 void MX_FREERTOS_Init(void) {
   /* USER CODE BEGIN Init */
-  stackmon_register("Main", mainTaskBuffer, sizeof(mainTaskBuffer));
-  stackmon_register("JobRunner", jobRunnerBuffer, sizeof(jobRunnerBuffer));
-  stackmon_register("Messaging", msgTaskBuffer, sizeof(msgTaskBuffer));
+  stackmon_register("Main", mainTaskStack, sizeof(mainTaskStack));
+  stackmon_register("JobRunner", jobRunnerStack, sizeof(jobRunnerStack));
+  stackmon_register("Messaging", msgTaskStack, sizeof(msgTaskStack));
   /* USER CODE END Init */
 
   /* Create the mutex(es) */
@@ -168,15 +168,15 @@ void MX_FREERTOS_Init(void) {
 
   /* Create the thread(s) */
   /* definition and creation of tskMain */
-  osThreadStaticDef(tskMain, TaskMain, osPriorityHigh, 0, TSK_STACK_MAIN, mainTaskBuffer, &mainTaskControlBlock);
+  osThreadStaticDef(tskMain, TaskMain, osPriorityHigh, 0, TSK_STACK_MAIN, mainTaskStack, &mainTaskControlBlock);
   tskMainHandle = osThreadCreate(osThread(tskMain), NULL);
 
   /* definition and creation of tskJobRunner */
-  osThreadStaticDef(tskJobRunner, TaskJobQueue, osPriorityAboveNormal, 0, TSK_STACK_JOBRUNNER, jobRunnerBuffer, &jobRunnerControlBlock);
+  osThreadStaticDef(tskJobRunner, TaskJobQueue, osPriorityAboveNormal, 0, TSK_STACK_JOBRUNNER, jobRunnerStack, &jobRunnerControlBlock);
   tskJobRunnerHandle = osThreadCreate(osThread(tskJobRunner), NULL);
 
   /* definition and creation of TaskMessaging */
-  osThreadStaticDef(tskMsg, TaskMessaging, osPriorityNormal, 0, TSK_STACK_MSG, msgTaskBuffer, &msgTaskControlBlock);
+  osThreadStaticDef(tskMsg, TaskMessaging, osPriorityNormal, 0, TSK_STACK_MSG, msgTaskStack, &msgTaskControlBlock);
   tskMsgHandle = osThreadCreate(osThread(tskMsg), NULL);
 
   /* USER CODE BEGIN RTOS_THREADS */
diff --git a/gex.mk b/gex.mk
index 7bb2042..f630515 100644
--- a/gex.mk
+++ b/gex.mk
@@ -83,8 +83,8 @@ GEX_CDEFS = $(GEX_CDEFS_BASE) \
     -DVERBOSE_ASSERT=1 \
     -DDEBUG_VFS=0 \
     -DDEBUG_FLASH_WRITE=0 \
-    -DVERBOSE_HARDFAULT=0 \
-    -DUSE_STACK_MONITOR=0 \
+    -DVERBOSE_HARDFAULT=1 \
+    -DUSE_STACK_MONITOR=1 \
     -DUSE_DEBUG_UART=1
 
 endif
diff --git a/platform/plat_compat.h b/platform/plat_compat.h
index c8e0627..e571af6 100644
--- a/platform/plat_compat.h
+++ b/platform/plat_compat.h
@@ -8,17 +8,23 @@
 #define VFS_DRIVE_NAME "GEX"
 
 // -------- Static buffers ---------
-#define TSK_STACK_MAIN      220 // USB / VFS task stack size
-#define TSK_STACK_MSG       220 // TF message handler task stack size
-#define TSK_STACK_JOBRUNNER 80 // Job runner task stack size
+// USB / VFS task stack size
+#if DISABLE_MSC
+ #define TSK_STACK_MAIN      100 // without MSC the stack usage is significantly lower
+#else
+ #define TSK_STACK_MAIN      160
+#endif
+
+#define TSK_STACK_MSG       220 // TF message handler task stack size (all unit commands run on this thread)
+#define TSK_STACK_JOBRUNNER 80 // Job runner task stack size (for async execution of events caught in interrupt)
 
 #define BULK_READ_BUF_LEN 256   // Buffer for TF bulk reads
 #define UNIT_TMP_LEN      512   // Buffer for bulk read and varions internal unit operations
 
-#define FLASH_SAVE_BUF_LEN  256 // Static buffer for saving to flash
+#define FLASH_SAVE_BUF_LEN  128 // Static buffer for saving to flash
 
 #define JOB_QUEUE_CAPACITY  4 // Job runner queue size (16 bytes each)
-#define RX_QUE_CAPACITY    10 // TinyFrame rx queue size (64 bytes each)
+#define RX_QUE_CAPACITY    6 // TinyFrame rx queue size (64 bytes each)
 
 #define TF_MAX_PAYLOAD_RX 512 // TF max Rx payload
 #define TF_SENDBUF_LEN     64 // TF transmit buffer (can be less than a full frame)