I should note; there are no problems with 20nm that I know of. The shuttle/full mask details I have seem to be standard timescales - which should have been known by all hardware manufacturers in the first instance prior to making any promises on shipping schedules.
Depends on what you constitute as "problems" I guess, first eta for 20nm gpus were late Q1/Q2 2014 around a year ago, Nvidia didn't just push maxwell on 20 nm into Q1 2015 for no good reason. It could simply be supply constraints and Apple pretty much buying up every single 20nm wafer leaving everyone else stranded.
granted, and as a total neutral in the matter, it does seem that most manufacturers have been a tad overzealous, in the promises they have made.