Every terminating Turing machine running in logarithmic space terminates in polynomial time. This is because the number of configurations in a logarithmic space Turing machine is polynomial; a logspace Turing machine running longer than the number of configurations will necessarily repeat a configuration, and so will never terminate.
More generally, $\mathsf{SPACE}(f(n)) \subseteq \mathsf{TIME}(2^{O(f(n))})$ for similar reasons. This gives not only $\mathsf{L} \subseteq \mathsf{P}$ but also $\mathsf{PSPACE} \subseteq \mathsf{EXPTIME}$.