Also, in case it wasn't clear from the previous post, the "much easier way" to solve the problems referred only to this particular scenario we're discussing with /dev/mem. It should be clear that the protection discussed in the paper is useful against kernel exploitation as well and in fact is quite close to the concept of KERNSEAL which the PaX team came up with in 2003. Google reveals a couple results on KERNSEAL, but the PaX team hasn't released a definitive explanation of the implementation, which is likely why it wasn't referenced by the paper you linked to.