Let's first look at a simpler scenario that one can have in a lab. Then afterwards, we can apply it to gravity.
When light (or a photon) reflects from a mirror, the change in propagation direction implies a change in the momentum of the photon. Momentum conservation requires that the mirror picked up the difference and therefore must acquire some small momentum kick from the photon. However, a photon produced in a lab will not be a perfect plane wave and therefore would consist of a superposition of multiple different plane waves with (slightly) difference momenta. When such a superposition reflects from the mirror, it will receive different kicks from the different plane waves in the superposition. Hence, it must become entangled with the photon. However, the effect is so small that it is generally safe to ignore it, as it is always done in the literature. Note, this has nothing to do with any mathematical formulation of the scenario and whether or not we can represent the mirror as a quantum state or not. It's just physics.
No apply this to a photon propagating past a planet. We know from physical observations that the gravity of the planet will change the propagation direction of the photon. Now we have the same situation as with the mirror in the lab: the photon must become entangled with the planet. Here the effect is even smaller than in the lab, and thus once again safely ignored.
Finally, we can consider the scenario that you describe. In principle, nothing prevents the photon from becoming entangled with gravity based on purely physical arguments. However, once again the effect would be so small that it would be extremely hard to measure it.