To me your choice of boundary conditions seems strange. I would not have a Dirichlet condition for both the pressure and velocity on the same boundary (see for instance: https://bitbucket.org/fenics-project/dolfin/src/master/python/demo/documented/navier-stokes/demo_navier-stokes.py).
If you choose to have only dirichlet conditons for the pressure, you need to either:
a) remove the pressure nullspace from the pressure update step.
b) Average the pressure after solving, as shown in: Order of convergence for a Navier-Stokes solver