it's actually not a bug, but something i'd like to investigate whenever i get some time. probably a debug option i left enabled somewhere
this was observed on a thinkpad x200 and several other machines
could be the raminit thing. it was observed on gm45 hardware, and coreboot raminit is known to silently fail on that (at the training stage) and retry a few times before succeeding.
in recent coreboot, this happens more often than in 2016 coreboot. we don't know why, or how to fix it yet. anyway, this could be the cause.