The mineralogy of the planetary interiors is critical in controlling the dynamics and evolution of super-Earths. For large super-Earths, the post-post-perovskite (post-pPv) phase Mg2SiO4, one of the major mantle phases, may undergo the order-disorder transition (ODT) at high temperatures. Yet, the ODT phase boundary of Mg2SiO4 has not been rigorously constrained. Additionally, fundamental thermodynamic properties of the disordered Mg2SiO4 remain to be determined. Here, we develop a unified machine-learning potential (MLP) for Mg2SiO4 of ab initio quality across super-Earth mantle conditions. This efficient MLP enables us to extensively calculate the free energy of post-pPv Mg2SiO4 using the thermodynamic integration method. The results are used to construct the ODT phase boundary. Furthermore, we report the equation of state and Grüneisen parameters for post-pPv Mg2SiO4 with various degrees of disorder. These thermodynamic properties are used to update the adiabatic thermal profiles and the mass-radius relation of super-Earths.