Fix 'get_attr' call in dynamo 'run_node' - #127696

======================================================================
ERROR: test_backend_with_compiler_delegate_and_operator_with_two_modules (executorch.exir.backend.test.test_backends.TestBackends) (extract_delegate_segments=False)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "<truncated path>/executorch/exir/tracer.py", line 665, in dynamo_trace
    return torchdynamo.export(
  File "<truncated path>/torch/_dynamo/eval_frame.py", line 1378, in inner
    result_traced = opt_f(*args, **kwargs)
  File "<truncated path>/torch/nn/modules/module.py", line 1552, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
  File "<truncated path>/torch/nn/modules/module.py", line 1561, in _call_impl
    return forward_call(*args, **kwargs)
  File "<truncated path>/torch/_dynamo/eval_frame.py", line 432, in _fn
    return fn(*args, **kwargs)
  File "<truncated path>/torch/nn/modules/module.py", line 1552, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
  File "<truncated path>/torch/nn/modules/module.py", line 1561, in _call_impl
    return forward_call(*args, **kwargs)
  File "<truncated path>/torch/_dynamo/convert_frame.py", line 1115, in __call__
    return self._torchdynamo_orig_c
...
...
...
<truncated path>/torch/_dynamo/symbolic_convert.py", line 742, in call_function
    self.push(fn.call_function(self, args, kwargs))
  File "<truncated path>/torch/_dynamo/variables/higher_order_ops.py", line 1151, in call_function
    real_sub_args = pytree.tree_map_only(
  File "<truncated path>/torch/utils/_pytree.py", line 1120, in tree_map_only
    return tree_map(map_only(__type_or_types_or_pred)(func), tree, is_leaf=is_leaf)
  File "<truncated path>/torch/utils/_pytree.py", line 948, in tree_map
    return treespec.unflatten(map(func, *flat_args))
  File "<truncated path>/torch/utils/_pytree.py", line 787, in unflatten
    leaves = list(leaves)
  File "<truncated path>/torch/utils/_pytree.py", line 1066, in wrapped
    return func(x)
  File "<truncated path>/torch/_dynamo/variables/higher_order_ops.py", line 1152, in <lambda>
    torch.fx.Proxy, lambda a: get_real_value(a.node, tx.output), p_args
  File "<truncated path>/torch/_dynamo/utils.py", line 1941, in get_real_value
    args, kwargs = torch.fx.node.map_arg(
  File "<truncated path>/torch/fx/node.py", line 744, in map_arg
    return map_aggregate(a, lambda x: fn(x) if isinstance(x, Node) else x)
  File "<truncated path>/torch/fx/node.py", line 752, in map_aggregate
    t = tuple(map_aggregate(elem, fn) for elem in a)
  File "<truncated path>/torch/fx/node.py", line 752, in <genexpr>
    t = tuple(map_aggregate(elem, fn) for elem in a)
  File "<truncated path>/torch/fx/node.py", line 752, in map_aggregate
    t = tuple(map_aggregate(elem, fn) for elem in a)
  File "<truncated path>/torch/fx/node.py", line 752, in <genexpr>
    t = tuple(map_aggregate(elem, fn) for elem in a)
  File "<truncated path>/torch/fx/node.py", line 762, in map_aggregate
    return fn(a)
  File "<truncated path>/torch/fx/node.py", line 744, in <lambda>
    return map_aggregate(a, lambda x: fn(x) if isinstance(x, Node) else x)
  File "<truncated path>/torch/_dynamo/utils.py", line 1943, in <lambda>
    lambda n: get_real_value(n, tracer),
  File "<truncated path>/torch/_dynamo/utils.py", line 1961, in get_real_value
    raise TorchRuntimeError(str(e)).with_traceback(e.__traceback__) from None
  File "<truncated path>/torch/_dynamo/utils.py", line 1958, in get_real_value
    real_value = run_node(tracer, node, args, kwargs, nn_module)
  File "<truncated path>/torch/_dynamo/utils.py", line 1922, in run_node
    raise RuntimeError(make_error_message(e)).with_traceback(
  File "<truncated path>/torch/_dynamo/utils.py", line 1911, in run_node
    return tracer.output_graph.get_submodule(node.target)
torch._dynamo.exc.TorchRuntimeError: Failed running get_attr L__self___lowerable(*(), **{}):
'OutputGraph' object has no attribute 'output_graph'

from user code:
   File "<truncated path>/executorch/exir/backend/test/test_backends.py", line 533, in forward
    b = self.lowerable(a)
  File "<truncated path>/torch/nn/modules/module.py", line 1561, in _call_impl
    return forward_call(*args, **kwargs)
  File "<truncated path>/executorch/exir/lowered_backend_module.py", line 336, in forward
    return executorch_call_delegate(self, *args)


The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "<truncated path>/executorch/exir/backend/test/test_backends.py", line 92, in wrapper
    test_method(self, extract_delegate_segments=extract_delegate_segments)
  File "<truncated path>/executorch/exir/backend/test/test_backends.py", line 546, in test_backend_with_compiler_delegate_and_operator_with_two_modules
    exir.capture(composite_model, model_inputs, exir.CaptureConfig())
  File "<truncated path>/executorch/exir/capture/_capture.py", line 232, in capture
    graph_module, _ = dynamo_trace(
  File "<truncated path>/executorch/exir/tracer.py", line 686, in dynamo_trace
    raise InternalError(
executorch.exir.error.InternalError: torchdynamo internal error occured. Please see above stacktrace