If one wants to get to the bottom of it, old and new factory images can be extracted via tricky application of squashfs tools (and maybe dd and whatever compression tool, I forget) and the various startup scripts and programs diffed directly. Or you can poke around on the running system (or two side by side?) and see what is changed. Of course checking any changes in the public repo would be worthwhile, but unclear if whatever is causing the issue made it there. My initial sense after looking at it is that the upstream change WereCatf mentioned would not apply to this hardware, but not 100% sure of that. It would not be at all surprising if the issue does turn to to be a similarly tiny detail breaking some part of the automated process.