As the above comments mention, van der Waals interactions are the primary types of intermolecular forces forces involved in the adsorption of any molecule onto a surface. Given that PDMS does not react with the glass surface chemically, it is pretty much certain that it is adhering through the combination of many weaker intermolecular forces (dipole/dipole, dipole/induced-dipole, and induced-dipole/induced-dipole interactions). You do not need covalent bonds for a molecule to adhere to the surface, given that there are enough weaker intermolecular forces holding it in place (covalent bond dissociation energies are on the order of hundreds of kcal/mol, whereas van der Waals interactions are an order to two orders of magnitude weaker). PDMS is a polymer, so given its length it should not be a problem to have many van der Waals interactions with the surface.
Indeed, if we consider the molecular structures of PDMS and glass (which is made from silica), we see that the Si-O bonds in PDMS and glass are permanent dipoles (oxygen is more electronegative than silicon). Therefore, dipole/dipole interactions are certainly at play. Given the rest of the structure, it is also safe to assume that dipole/induced-dipole interactions will be present, as well as induced-dipole/induced-dipole interactions.
However, from a quick search I learned that glass is not only composed of silicon and oxygen. Glass can also contain additional species such as calcium and sodium ions which are the result of treating silica with sodium carbonate and/or calcium carbonate during the manufacturing process. Compounds such as these are typically added to lower the melting temperature of silica as well as to keep the finished glass from deteriorating (source: http://www.cmog.org/article/what-is-glass). Given that there are these additional ions (and perhaps others) present in glass, it follows that ion/dipole and ion/induced-dipole interactions are also contributing to the adsorption of PDMS onto the glass surface, in addition to van der Waals forces.
Because these forces are directional, I suspect that PDMS adsorbs onto the glass such that the Si-O bonds are facing the glass surface. This way, it may maximize its interaction with the Si-O dipoles in the glass, as well as the calcium and sodium cations. As a result, the less polar methyl groups would be left pointing outwards from the surface, giving the coated surface its hydrophobic properties.