See libhal/async_context:README.md#Avoid Operation Stacking for what operation stacking is.
When we create a top level coroutine we know that initial suspend is called and the coroutine does not proceed. The question is, does this only happen to top level coroutines? Does this step get bypassed if you co_await the future. If it is the case that initial_suspend is skipped when a future is co_awaited, then we can use this to check if the context's memory usage is non-zero. If its non-zero then we know that its preloaded.
But this assumes that somehow co_awaiting a future changes what happens when the coroutine's factory function is called. I'm not sure if there is such an asymmetry.