If test-and-set! is implemented as an ordinary procedure, two process can
acquire the mutex at the same time in the following way:
- Process P1 starts executing acquire and reaches the point where test-and-set!
  sees the cell is false and is interrupted by P2 before it can evaluate the
  else expression and set the cell to true.
- Process P2 interrupts P1 and acquires the mutex because P1 has not cell the
  cell to true yet.
- Process P1 resumes execution and sets the cell to true event though it already
  is true.
- Now both processes think they own the mutex.