As pointed out by Guennadi Liakhovetski (thanks), pin2 is already shifted
left by one. So the additional shift is bogus.
Signed-off-by: NStefan Roese <sr@denx.de>