This is what my book says , but consider this :
i take something in an adiabatic container with a piston and pull up the piston to change its height by [tex]\Delta[/tex]h in 1 step, then again I bring it back to same height in 1 step , now this process is not quasi-static but it has been reversed !!!
also it is back in the same state!!!

also why should the system be always in equilibrium to be reversable

The process will have to be accurate by 100%, that means the amount of work done on the system should be identical to the smallest amount of work done by the system...if you're measuring this, it's only achievable if you're observing the system by infinitely small steps, or making progress through infinitely small steps and each step is observed.

A smooth graph formed off the adiabatic process will mean that you have every infinity small detail proving that the area of the graph formed when work is done on the system = the area of the graph formed when work is done by the system...this is impossible in real life.