It really depends on the amount of time that is available. If there is not enough time, using the list of requirements available, identify at least two test cases (1 positive and 1 negative scenario) for each requirement. If time permits, use more structured techniques like Equivalence Class Partitioning and Boundary Value Analysis to identify test cases for each requirement. After identifying the test cases, try to normalize them to get the number of adequate test cases to test a software module.