Stefan Reinauer wrote:
Please do not accept "it seems to work" for new register descriptions,
I figured nobody has the time to check this, not even the author of msrtool.
Someone else using GX2 might hav been motivated (though they didn't speak up so far) or after the reminder maybe someone completely without experience in coreboot could use this as a way to get involved a little bit more. I tried to encourage this, but I could probably have been more clear.
No doubt that it's difficult to have good quality reviews happen quickly enough when not many people seem to be interested.
Should someone ever use this code on a Geode GX2 system again, the code is there and can be improved now.
My point is that users of the tools should not have to start by reviewing them. The tools are much less useful if we developers haven't tried our best to make them correct. Especially something brain-damaging as copying register descriptions from a data sheet can really benefit from reviews.
That's much better than annoying contributors and letting contributions going down the drain.
This is also a good point. Thanks. The balance between strict code review and accomodating contributions+contributors is a hard one.
Might it be useful to have a please-review branch, in addition to keeping track of submitted patches in patchwork?
//Peter