You reported that before, but unfortunately I don’t have a 4×4 to test. I know that the memory size detection in u-boot is somewhat flaky. Jon N. changed something a while ago but I heard bad things (i.e. some non-booting units) about this change. So I did not include it. However, I’m also not 100% sure if our kernel config has something to do with this. There is one parameter that looks suspicious… Maybe next week I can build a kernel with that param changed, so that we can rule that out.