py/emitnative: Implement yield and yield-from in native emitter.

This commit adds first class support for yield and yield-from in the native
emitter, including send and throw support, and yields enclosed in exception
handlers (which requires pulling down the NLR stack before yielding, then
rebuilding it when resuming).

This has been fully tested and is working on unix x86 and x86-64, and
stm32.  Also basic tests have been done with the esp8266 port.  Performance
of existing native code is unchanged.

py/objgenerator.c

@@ -73,6 +73,53 @@ const mp_obj_type_t mp_type_gen_wrap = {
     #endif
 };
 
+/******************************************************************************/
+// native generator wrapper
+
+#if MICROPY_EMIT_NATIVE
+
+STATIC mp_obj_t native_gen_wrap_call(mp_obj_t self_in, size_t n_args, size_t n_kw, const mp_obj_t *args) {
+    // The state for a native generating function is held in the same struct as a bytecode function
+    mp_obj_fun_bc_t *self_fun = MP_OBJ_TO_PTR(self_in);
+
+    // Determine start of prelude, and extract n_state from it
+    uintptr_t prelude_offset = ((uintptr_t*)self_fun->bytecode)[0];
+    size_t n_state = mp_decode_uint_value(self_fun->bytecode + prelude_offset);
+    size_t n_exc_stack = 0;
+
+    // Allocate the generator object, with room for local stack and exception stack
+    mp_obj_gen_instance_t *o = m_new_obj_var(mp_obj_gen_instance_t, byte,
+        n_state * sizeof(mp_obj_t) + n_exc_stack * sizeof(mp_exc_stack_t));
+    o->base.type = &mp_type_gen_instance;
+
+    // Parse the input arguments and set up the code state
+    o->globals = self_fun->globals;
+    o->code_state.fun_bc = self_fun;
+    o->code_state.ip = (const byte*)prelude_offset;
+    mp_setup_code_state(&o->code_state, n_args, n_kw, args);
+
+    // Indicate we are a native function, which doesn't use this variable
+    o->code_state.exc_sp = NULL;
+
+    // Prepare the generator instance for execution
+    uintptr_t start_offset = ((uintptr_t*)self_fun->bytecode)[1];
+    o->code_state.ip = MICROPY_MAKE_POINTER_CALLABLE((void*)(self_fun->bytecode + start_offset));
+
+    return MP_OBJ_FROM_PTR(o);
+}
+
+const mp_obj_type_t mp_type_native_gen_wrap = {
+    { &mp_type_type },
+    .name = MP_QSTR_generator,
+    .call = native_gen_wrap_call,
+    .unary_op = mp_generic_unary_op,
+    #if MICROPY_PY_FUNCTION_ATTRS
+    .attr = mp_obj_fun_bc_attr,
+    #endif
+};
+
+#endif // MICROPY_EMIT_NATIVE
+
 /******************************************************************************/
 /* generator instance                                                         */
 
@@ -118,7 +165,22 @@ mp_vm_return_kind_t mp_obj_gen_resume(mp_obj_t self_in, mp_obj_t send_value, mp_
     self->code_state.old_globals = mp_globals_get();
     mp_globals_set(self->globals);
     self->globals = NULL;
-    mp_vm_return_kind_t ret_kind = mp_execute_bytecode(&self->code_state, throw_value);
+
+    mp_vm_return_kind_t ret_kind;
+
+    #if MICROPY_EMIT_NATIVE
+    if (self->code_state.exc_sp == NULL) {
+        // A native generator, with entry point 2 words into the "bytecode" pointer
+        typedef uintptr_t (*mp_fun_native_gen_t)(void*, mp_obj_t);
+        mp_fun_native_gen_t fun = MICROPY_MAKE_POINTER_CALLABLE((const void*)(self->code_state.fun_bc->bytecode + 2 * sizeof(uintptr_t)));
+        ret_kind = fun((void*)&self->code_state, throw_value);
+    } else
+    #endif
+    {
+        // A bytecode generator
+        ret_kind = mp_execute_bytecode(&self->code_state, throw_value);
+    }
+
     self->globals = mp_globals_get();
     mp_globals_set(self->code_state.old_globals);