There's no reasonable purpose of a pull-up resistor at LD. I guess, you didn't think about it thoroughly. A pull-down resistor without concurrent current sources will allow an LM358 output voltage effectively down to zero (< 1 mV).
P.S.: Your output resistance calculations are missing the point. The datasheet clearly shows about 800 mV residual voltage with 1 mA sink current. If you want to understand why, you should review the (simplified) transistor level circuit in the datasheet. Above 50 µA sink current, the output voltage rises to one diode voltage drop. Below 50 µA, it's essentially the saturation voltage of a current mirror. With a pull-down resistor, it's zero plus leakage currents.
I think the erroneous pull-up resistor concept has been elaborated sufficiently now...
By the way, the pull-up resistor won't ever serve the intended purpose, it's always overriden by the LM358.