py: Add option to reduce GC stack integer size to save RAM.

A new option MICROPY_GC_STACK_ENTRY_TYPE is added to select a custom type
instead of size_t for the gc_stack array items.  This can be beneficial for
small devices, especially those that are low on memory anyway.  If a device
has 1MB or less of heap (and 16-byte GC blocks) then this type can be
uint16_t, saving 128 bytes of RAM.

py/mpconfig.h

@@ -106,6 +106,15 @@
 #define MICROPY_ALLOC_GC_STACK_SIZE (64)
 #endif
 
+// The C-type to use for entries in the GC stack.  By default it allows the
+// heap to be as large as the address space, but the bit-width of this type can
+// be reduced to save memory when the heap is small enough.  The type must be
+// big enough to index all blocks in the heap, which is set by
+// heap-size-in-bytes / MICROPY_BYTES_PER_GC_BLOCK.
+#ifndef MICROPY_GC_STACK_ENTRY_TYPE
+#define MICROPY_GC_STACK_ENTRY_TYPE size_t
+#endif
+
 // Be conservative and always clear to zero newly (re)allocated memory in the GC.
 // This helps eliminate stray pointers that hold on to memory that's no longer
 // used.  It decreases performance due to unnecessary memory clearing.

py/mpstate.h

@@ -77,7 +77,7 @@ typedef struct _mp_state_mem_t {
     byte *gc_pool_end;
 
     int gc_stack_overflow;
-    size_t gc_stack[MICROPY_ALLOC_GC_STACK_SIZE];
+    MICROPY_GC_STACK_ENTRY_TYPE gc_stack[MICROPY_ALLOC_GC_STACK_SIZE];
     uint16_t gc_lock_depth;
 
     // This variable controls auto garbage collection.  If set to 0 then the