Refine iMX RT memory layout and add three boards

Introduces a way to place CircuitPython code and data into
tightly coupled memory (TCM) which is accessible by the CPU in a
single cycle. It also frees up room in the corresponding cache for
intermittent data. Loading from external flash is slow!

The data cache is also now enabled.

Adds support for the iMX RT 1021 chip. Adds three new boards:
* iMX RT 1020 EVK
* iMX RT 1060 EVK
* Teensy 4.0

Related to #2492, #2472 and #2477. Fixes #2475.

6 files changed, 22 insertions, 6 deletions

diff --git a/supervisor/shared/filesystem.c b/supervisor/shared/filesystem.c
index 92727f6fb..3b799619f 100644
--- a/supervisor/shared/filesystem.c
+++ b/supervisor/shared/filesystem.c
@@ -75,7 +75,7 @@ static void make_sample_code_file(FATFS *fatfs) {
     FIL fs;
     UINT char_written = 0;
     const byte buffer[] = "print('Hello World!')\n";
-    //Create or modify existing code.py file 
+    //Create or modify existing code.py file
     f_open(fatfs, &fs, "/code.py", FA_WRITE | FA_CREATE_ALWAYS);
     f_write(&fs, buffer, sizeof(buffer) - 1, &char_written);
     f_close(&fs);
diff --git a/supervisor/shared/memory.c b/supervisor/shared/memory.c
index 38040d11d..14c3b4979 100755
--- a/supervisor/shared/memory.c
+++ b/supervisor/shared/memory.c
@@ -39,11 +39,14 @@ uint32_t* low_address;
 uint32_t* high_address;
 
 void memory_init(void) {
-    low_address = port_stack_get_limit();
-    high_address = port_stack_get_top();
+    low_address = port_heap_get_bottom();
+    high_address = port_heap_get_top();
 }
 
 void free_memory(supervisor_allocation* allocation) {
+    if (allocation == NULL) {
+        return;
+    }
     int32_t index = 0;
     bool found = false;
     for (index = 0; index < CIRCUITPY_SUPERVISOR_ALLOC_COUNT; index++) {
diff --git a/supervisor/shared/safe_mode.c b/supervisor/shared/safe_mode.c
index aba31e9c9..c957aee53 100644
--- a/supervisor/shared/safe_mode.c
+++ b/supervisor/shared/safe_mode.c
@@ -157,6 +157,9 @@ void print_safe_mode_message(safe_mode_t reason) {
             case FLASH_WRITE_FAIL:
                 serial_write_compressed(translate("Failed to write internal flash."));
                 break;
+            case MEM_MANAGE:
+                serial_write_compressed(translate("Invalid memory access."));
+                break;
             default:
                 serial_write_compressed(translate("Unknown reason."));
                 break;
diff --git a/supervisor/shared/safe_mode.h b/supervisor/shared/safe_mode.h
index e05fca0e4..5b09c4b54 100644
--- a/supervisor/shared/safe_mode.h
+++ b/supervisor/shared/safe_mode.h
@@ -40,6 +40,7 @@ typedef enum {
   PROGRAMMATIC_SAFE_MODE,
   NORDIC_SOFT_DEVICE_ASSERT,
   FLASH_WRITE_FAIL,
+  MEM_MANAGE,
 } safe_mode_t;
 
 safe_mode_t wait_for_safe_mode_reset(void);
diff --git a/supervisor/shared/stack.c b/supervisor/shared/stack.c
index dcecf2067..2b7b1c03a 100755
--- a/supervisor/shared/stack.c
+++ b/supervisor/shared/stack.c
@@ -46,6 +46,10 @@ void allocate_stack(void) {
 
     mp_uint_t c_size = (uint32_t) port_stack_get_top() - sp;
 
+    if (port_stack_get_top() != port_heap_get_top()) {
+        return;
+    }
+
     stack_alloc = allocate_memory(c_size + next_stack_size + EXCEPTION_STACK_SIZE, true);
     if (stack_alloc == NULL) {
         stack_alloc = allocate_memory(c_size + CIRCUITPY_DEFAULT_STACK_SIZE + EXCEPTION_STACK_SIZE, true);
@@ -71,6 +75,9 @@ void stack_init(void) {
 }
 
 void stack_resize(void) {
+    if (stack_alloc == NULL) {
+        return;
+    }
     if (next_stack_size == current_stack_size) {
         *stack_alloc->ptr = STACK_CANARY_VALUE;
         return;
diff --git a/supervisor/shared/tick.c b/supervisor/shared/tick.c
index 69256081c..5668f8fa1 100644
--- a/supervisor/shared/tick.c
+++ b/supervisor/shared/tick.c
@@ -25,11 +25,13 @@
  */
 
 #include "supervisor/shared/tick.h"
+
+#include "supervisor/linker.h"
 #include "supervisor/filesystem.h"
 #include "supervisor/shared/autoreload.h"
 
-static volatile uint64_t ticks_ms;
-static volatile uint32_t background_ticks_ms32;
+static volatile uint64_t PLACE_IN_DTCM_BSS(ticks_ms);
+static volatile uint32_t PLACE_IN_DTCM_BSS(background_ticks_ms32);
 
 #if CIRCUITPY_GAMEPAD
 #include "shared-module/gamepad/__init__.h"
@@ -77,7 +79,7 @@ uint32_t supervisor_ticks_ms32() {
 
 extern void run_background_tasks(void);
 
-void supervisor_run_background_tasks_if_tick() {
+void PLACE_IN_ITCM(supervisor_run_background_tasks_if_tick)() {
     uint32_t now32 = ticks_ms;
 
     if (now32 == background_ticks_ms32) {