Downforce or lift only comes as a result of a differential in pressure.
You cannot consider the local diffuser without considering the pressure conditions on the upper surface as well.
As for the pressure conditions under the diffuser, local boundary layer effects cause the low pressure zones downstream of the ramp point. The effective area change doesn't kick in until the boundary layer is following the shape of the diffuser, and not in a small separation bubble at the ramp itself.
In an inviscid case, they wouldn't occur, flow would start to decelerate instantly as a function of the conservation of mass relationship.
(Assuming density is fixed, velocity will reduce linearly with increased area)
At the trailing edge of the diffuser, Kutta condition is enforced - there will be no downforce created at the trailing edge. The flow pressures will be equal top & bottom at the diffuser trailing edge.
There may be a shear layer between top and bottom flows, infact, it is probably advantageous to have one to try and entrain (drag/pull by momentum exchanges) more air out from under the floor.
The rake angle roger44 refers to is simply to increase the difference between the effective front area of the floor, and the effective rear area of the floor = more massflow rate under the car = more downforce.