Regardless whether you define that the mirror is moving toward the laser or laser toward the light the light beam always moves relative to both at c. So what creates this extra energy?
The first thing you need to realize is that the total energy you measure in a system is a property of your frame of reference. Total energy is kind of arbitrary -- energy's always measured WRT some baseline energy level. All inertial observers will agree that the total energy of a closed system does not change. But they way they think that energy is partitioned between parts of that system may vary. Non-inertial observes may observe energy is not conserved.
To make things simple, I'm going to make the mirror a freely moving inertial object.
In the laser's inertial reference frame (the laser is stationary in that frame), the energy increase in the beam comes from a retardation of the mirror's motion. Note there's no change in the total energy of this system, in this reference frame (as expected).
Now, note that in this set up, the mirror experiences an acceleration, and therefore it is not an inertial observer. We can consider two different inertial reference frames -- that of the mirror before reflection and that of the mirror after reflection.
Let's start with the before-reflection reference frame. The light beam comes from one direction and then bounces off to return back, and the mirror experiences a recoil. After the collision the mirror has some kinetic energy, so the reflected beam must have a lower energy than before -- it will be red shifted. The laser has the same energy as before, as its velocity is unchanged wrt the before-reflection reference frame.
In the after-reflection reference frame, the mirror comes from one direction and the light beam from the other. They meet, the mirror stops and the light beam returns from whence it came. In this case, the mirror loses energy, so the outgoing beam must have a higher energy -- it will be blue shifted.
How can the beam be both blue and red shifted? Well the reference frames we observe these phenomena in are moving relative to each other, and that accounts for it. Those observing red shift are moving in the direction of the reflected beam faster than those observing blue shift are moving in that direction.