Currently, the test uses hard-coded 0 as the return value of efa_mock_ibv_read_qp_num_return_mock, which is wrong as the qpn can be non-zero when there are other qp that share the device. This patch fixes this issue by using the correct qp num associated with the ep.
Currently, the test uses hard-coded 0 as the return value of efa_mock_ibv_read_qp_num_return_mock, which is wrong as the qpn can be non-zero when there are other qp that share the device. This patch fixes this issue by using the correct qp num associated with the ep.