PULP divsqrt unit does not produce correct results and stalls the unit with busy signal high

PULP Divsqrt unit is not working for all FP formats including FP8. Tried with different configurations (no pipeline/ w pipeline regs). Other divsqrt units works but I need to have FP8 divsqrt unit. This stalls the fp unit(Busy) and does not produce a result at any latancy. attached image is for FP32