Let $k$ be a nonarchimedean locally compact field, $G$ a connected reductive group over $k$ with minimal parabolic subgroup $P = MN$. Let $\sigma$ be a smooth representation of $M$. In chapter 3 of Casselman's notes on representation theory, he writes "If $P$ is a minimal parabolic subgroup and $\sigma$ is irreducible (hence necessarily finite dimensional)..."
Why does $P$ being a minimal parabolic imply that $\sigma$ is finite dimensional? This is clear if $P$ is a Borel subgroup, because then $M$ is commutative, making $\sigma$ one dimensional. In general, $M$ is a reductive group. This seems to indicate that there are constraints on which reductive groups can occur as Levi subgroups of minimal parabolic subgroups..
Edit: Let $S$ be a maximal split torus of $G$ which is contained in a minimal parabolic $P$. Then the set of roots of $S$ in $M$ is empty. So the question comes down to the following:
If $G$ is a connected, reductive group over $k$ with maximal split torus $S$, and $S$ is contained in the center of $G$, then every smooth irreducible representation of $G$ is finite dimensional.