Priority lock for multiprocessing shared memory

I have this code

def writer_entrypoint(shared_data: Array):
    while True:
        ind, new_value = get_data() # It takes up to 100 milliseconds

        with shared_data.get_lock():
             shared_data[ind] = new_value

def reader_entrypoint(shared_data: Value):
    while True:
        with shared_data.get_lock():
             data = np.frombuffer(shared_data.get_obj()).copy()
        # Process the data... but faster

Both functions are processes entry points. There are multiple readers and a single writer. The readers are much faster than the writer. And the writer waits a lot of time for the lock (starving)

So the reader uses "old data" but the writer can't update the data.

I want the writer will have a higher priority for updating the data. I mean, if there is a reader that uses the lock, wait for it. But when it finishes and the writer and several readers wait for the lock, let the writer use it first.

How can I implement such a mechanism?

source https://stackoverflow.com/questions/71067784/priority-lock-for-multiprocessing-shared-memory