diff --git a/arch/arm64/src/common/arm64_cpustart.c b/arch/arm64/src/common/arm64_cpustart.c
index 2ca32ea986c..2492ee3612e 100644
--- a/arch/arm64/src/common/arm64_cpustart.c
+++ b/arch/arm64/src/common/arm64_cpustart.c
@@ -222,15 +222,6 @@ int up_cpu_start(int cpu)
   sched_note_cpu_start(this_task(), cpu);
 #endif
 
-#ifdef CONFIG_STACK_COLORATION
-  /* If stack debug is enabled, then fill the stack with a
-   * recognizable value that we can use later to test for high
-   * water marks.
-   */
-
-  arm64_stack_color(g_cpu_idlestackalloc[cpu], SMP_STACK_SIZE);
-#endif
-
   cpu_boot_params.cpu_ready_flag = 0;
   arm64_start_cpu(cpu, (char *)g_cpu_idlestackalloc[cpu], SMP_STACK_SIZE,
                   arm64_smp_init_top);
diff --git a/arch/arm64/src/common/arm64_head.S b/arch/arm64/src/common/arm64_head.S
index 60ca2a65c63..357fefa4104 100644
--- a/arch/arm64/src/common/arm64_head.S
+++ b/arch/arm64/src/common/arm64_head.S
@@ -26,6 +26,7 @@
 #include <nuttx/config.h>
 
 #include "arm64_arch.h"
+#include "arm64_internal.h"
 #include "arm64_macro.inc"
 
 /****************************************************************************
@@ -153,7 +154,22 @@ real_start:
     /* we can now load our stack pointer value and move on */
 
     ldr    x24, [x0, #BOOT_PARAM_SP]
-    add    x24, x24, #(CONFIG_IDLETHREAD_STACKSIZE)
+
+#  ifdef CONFIG_STACK_COLORATION
+    /* Write a known value to the IDLE thread stack to support stack
+     * monitoring logic
+     */
+
+    ldr    w1, =SMP_STACK_WORDS
+    ldr    w2, =STACK_COLOR
+
+.loop:
+    sub    w1, w1, #1
+    str    w2, [x24], #4
+    cmp    w1, #0
+    bne    .loop
+#  endif
+
     ldr    x25, =arm64_boot_secondary_c_routine
     bl     __reset_prep_c
 
diff --git a/arch/arm64/src/common/arm64_internal.h b/arch/arm64/src/common/arm64_internal.h
index 5895ee41192..6ad103d75ee 100644
--- a/arch/arm64/src/common/arm64_internal.h
+++ b/arch/arm64/src/common/arm64_internal.h
@@ -91,6 +91,25 @@
 #define STACK_COLOR    0xdeaddead
 #define HEAP_COLOR     'h'
 
+/* AArch64 the stack-pointer must be 128-bit aligned */
+
+#define STACK_ALIGNMENT     16
+
+/* Stack alignment macros */
+
+#define STACK_ALIGN_MASK    (STACK_ALIGNMENT - 1)
+#define STACK_ALIGN_DOWN(a) ((a) & ~STACK_ALIGN_MASK)
+#define STACK_ALIGN_UP(a)   (((a) + STACK_ALIGN_MASK) & ~STACK_ALIGN_MASK)
+
+#ifdef CONFIG_SMP
+/* The size of interrupt and idle stack.  This is the configured
+ * value aligned the 8-bytes as required by the ARM EABI.
+ */
+
+#  define SMP_STACK_SIZE    STACK_ALIGN_UP(CONFIG_IDLETHREAD_STACKSIZE)
+#  define SMP_STACK_WORDS   (SMP_STACK_SIZE >> 2)
+#endif
+
 /****************************************************************************
  * Public Types
  ****************************************************************************/
@@ -112,16 +131,6 @@ extern "C"
 #define EXTERN extern
 #endif
 
-/* AArch64 the stack-pointer must be 128-bit aligned */
-
-#define STACK_ALIGNMENT     16
-
-/* Stack alignment macros */
-
-#define STACK_ALIGN_MASK    (STACK_ALIGNMENT - 1)
-#define STACK_ALIGN_DOWN(a) ((a) & ~STACK_ALIGN_MASK)
-#define STACK_ALIGN_UP(a)   (((a) + STACK_ALIGN_MASK) & ~STACK_ALIGN_MASK)
-
 #define INIT_STACK_DEFINE(sym, size) \
     char locate_data(".initstack") \
      aligned_data(STACK_ALIGNMENT) sym[size]
@@ -143,12 +152,6 @@ extern "C"
 
 #ifdef CONFIG_SMP
 
-/* The size of interrupt and idle stack.  This is the configured
- * value aligned the 8-bytes as required by the ARM EABI.
- */
-
-#define SMP_STACK_SIZE       STACK_ALIGN_UP(CONFIG_IDLETHREAD_STACKSIZE)
-
 INIT_STACK_ARRAY_DEFINE_EXTERN(g_cpu_idlestackalloc, CONFIG_SMP_NCPUS,
                           SMP_STACK_SIZE);
 INIT_STACK_ARRAY_DEFINE_EXTERN(g_interrupt_stacks, CONFIG_SMP_NCPUS,