Since the aligned allocation of shader closures in OSL memory pool this workaround is no longer needed. Also put a comment which describes the desired layout of the structure so array of shader closures is all nicely aligned.