Finding a two-sided confidence interval

Question

Let $X_1, \dotsc, X_n$ be i.i.d. $N(\mu,\sigma^2)$, where $\sigma^2$ is known, but $\mu$ is unknown. Construct a two sided 95% CI for $\mu$ based on $X_1, \dotsc, X_n$?

I am unsure how to go about this problem when there is an unknown involved.