I haven't done any in depth analysis of gryadka but I think the premise of the argument here may be wrong. Even if a particular value isn't accepted by a majority of nodes, that doesn't necessarily make it a dirty read. As long as anyone participating in the algorithm sees the history as if it was committed right before the cas result is committed, it could still be linearizable. I would need to create a formal model to be sure whether it is correct or not, but don't just assume it's wrong because it isn't paxos (which it isn't and it shouldn't advertise like it is).
But depending on node failures, couldn't the same client successfully run cas(nil, A) -> cas(A, B) -> cas(nil, C), with all operations succeeding? Say the first two operations only succeed in writing to a single node (as in the post's example). Then if that single node goes down, the third cas will succeed, which is certainly not linearizable.
Note: I'm working of the assumption that the system is correctly described by the posts author. The original author of Gryadka has disputed the description in the post, and I haven't read the source.
