Nico Huber has posted comments on this change. ( https://review.coreboot.org/c/coreboot/+/42455 )
Change subject: nb/intel/sandybridge/gma.c: Correct deasserting force wake on IVB ......................................................................
Patch Set 2:
I know it's a lot of work, but I think all the changes in this series (even individual bits) should be tested separately if they fix anything.
It's quite difficult to do given that we have no idea what these bits mean. Or by testing you mean running it for a couple of days and making sure the system is still stable?
Basically yes, do changes step by step until your problematic system is stable. If we merge this without detailed information which parts actually fix something, and then something goes wrong, it would be reverted as a whole.
Okay. But if in the end I'll find out that my crashes are fixed by one or two or three lock bits (not all four) moved to the end, and the reference code sets them all in the end (which actually makes sense to me), I hope we agree that we should move all four locks to the end?
I honestly don't care what we do as long as it is properly documented.
OTOH, if it turns out that all register writes are effective even if we don't move the locking, I see no reason to move it (assume the registers can be dumped).