Thinking a bit about it. We have three conditions.
First, the engine is driving the prop, the prop is using engine power to accelerate air and thus produce thrust. The air is moving over the prop airfoil in a controlled manner, in this case it is the way it was designed to work so fairly efficiently.
Second, the engine is producing no power and the prop is windmilling. The ier flowing through the prop is having energy extracted from it to provide the power to turn the dead engine. There would still be some sort of controlled airflow over the prop airfoil for this to work although the whole process would be pretty inefficient. I suspect that a windmill airfoil would look a lot different from a prop airfoil. But in any case drag is produced.
Third case, the prop simply produces drag and, as someone pointed out earlier, some torque which has to be countered by aileron at the cost of more drag. It's just another bit of profile drag from the point of view of aircraft performance.
From the various comments it would seem that there is not a lot of real world difference between the windmilling condition and the prop stopped condition.
Now the thing that I am wondering about is this. Could it be that in the case of an airplane which glides best at a low IAS, which would be a high aspect machine, might the stopped prop produce less drag? Conversely, in the case of an aircraft which glides best going faster, low aspect ration, might the windmilling prop produce less drag?