To answer this question, it's important to understand that Kant distinguishes sharply between morality and legality. The most commonly read Kant text on ethics Groundwork of the Metaphysics of Morals is singularly focused on morality. Kant's ethics is further spelled in the Critique of Practical Reason and the second half of the Metaphysics of Morals, the Tugendlehre which is often called "Doctrine of Virtue."
Kant's primary treatment of law occurs in the first section of his longer Metaphysics of Morals, the Rechtslehre which is called "Doctrine of Right" or "Metaphysical Principles of Justice."
This is important because Kant thinks that ethics is about what the reasons for your actions but law cannot look into these actions (Kant explains this about ethics by itself in several of these texts but spells it out for the distinction in the preface to MPJ).
This has several consequences. First, we cannot externally know whether someone's action is moral. At best we can only know if their action is compatible with morality. This is best explained by illustration, let's say the moral thing to do is to help an old lady across the street because she requires assistance navigating a busy intersection. Many different actions will involve me "helping the old lady across the street" but the action is only moral (on Kant's view) if I also got the reasons behind my action right.
Second, this applies on Kant's view not just to evaluating other's actions but quite possibly to our own (this ties back interestingly to Kant's own pietistic background and the belief we are all sinners). Consequently, laws do not extend to motives.
With one further piece of information, all of the pieces are on the board to answer your question. For Kant, we are prima facie to assume the rationality of others in their activities. This includes their law-making. Consequently, we are to assume even in the face of evidence to the contrary that laws that are promulgated are just.
So then to simply answer your question, Kant's position that a law that includes provisions for its own disobedience is self-contradictory is contradictory on a level far prior to the problem of an unjust law. For Kant, the very definition of law is something that it is right to obey (here right in the sense of justice not necessarily ethics).
In answer to your question about how would Kant handle unjust laws. Textual references escape me a bit but the basic answer is that your obligations to right (justice) are subservient to your obligations to ethics but only to the extent that they directly contradict.
When I stated above that Kant's ethics is about why we act and that we cannot know we acted morally (since it's not clear we have reflective access to our maxims on a sufficiently clear level), there's an interesting non-symmetry: sometimes we can know that an action is wrong. So since there are some actions that the categorical imperative (at least on Kant's own reading) expressly prohibits such as lying, murdering, we can know that any law commanding us to do these actions is a law we cannot obey... but it's important to realize that this is narrowly construed down to only the point of doing something that must be immoral because it's incompatible with the CI.