Work grumble
Of course, after months and months and months of using certain methods working flawlessly to program and boot our prototype boards. We get the first production builds of the products, annnnd... they fail to boot after programming. Good job, me, whatever I did.
Hopefully I can figure out what the issue is and recover the boards, but this stuff is danged hard to debug. *sigh*
Woohoo! Turns out I was able to recover the boards with a little "manual intervention", but the workaround is really weird. I basically run a certain step that sets the necessary magic bits twice. The first time (done automatically) doesn't take, but the second time (done manually by me) does take. I still have no idea why this steps suddenly stopped working, but since I recovered the boards, I can now get some sleep this weekend.
