In this paper, the downlink of a multi-cell massive MIMO system is considered where the channel state information (CSI) is estimated via pilot symbols that are orthogonal in a cell but re-used in other cells. Re-using the pilots, however, contaminates the CSI estimate at each base station (BS) by the channel of the users sharing the same pilot in other cells. The resulting inter-cell interference does not vanish even when the number of BS antennas $M$ is large, i.e., $M\rightarrow\infty$, and thus the rates achieved by treating interference as noise (TIN) saturate even if $M\rightarrow\infty$. In this paper, interference aware decoding schemes based on simultaneous unique decoding (SD) and simultaneous non-unique decoding (SND) of the full interference or a part of the interference (PD) are studied with two different linear precoding techniques: maximum ratio transmission (MRT) and zero forcing (ZF). The resulting rates are shown to grow unbounded as $M\rightarrow\infty$. In addition, the rates achievable via SD/SND/PD for finite $M$ are derived using a worst-case uncorrelated noise technique, which are shown to scale as $\mathcal{O}(\log M)$. To compare the performance of different schemes, the maximum symmetric rate problem is studied, where it is confirmed that with large, yet practical, values of $M$, SND strictly outperforms TIN, and also that PD strictly outperforms SND.