On Fri, 04 Jun 2021, Wolfram Sang wrote: > > > IMHO, we wouldn't want to foster the impression that it's okay to > > provide a non-determined effort, safe in the knowledge that someone > > will come along later and finish the job for them at a later date. > > Right. > > The first lesson from that is that maintainers should require > documentation of the fields when they get added. This was my oversight > because it was back then not reported by checkers, probably. I am sorry. > It annoys me, too. Sure. When I started this work, there were 18k+ W=1 warnings in the kernel. Now there are more like 3k. I don't think anyone is to blame per say, it's just something that people haven't particularly cared about up until this point. One of my main aims is to clean-up W=1s to the point where enabling them would become normal practice, rather than the situation we're in presently where enabling them just dominates the build-log, making them more trouble than they're worth. > If I notice that someone updates a driver which doc-errors, then I ask > if that could be fixed by this person, too. It usually works. Not for > drivers without attention, of course. But this is why I don't mind > doc-errors to stay. I'd rather they didn't say. This would void the main aim of this effort. > If this is considered problematic, then I'd suggest to remove the kernel > doc headers like you did, but add a comment like: > > * FIXME: add missing fields and reenable kernel-doc > > To make sure, it is obvious even by glimpsing through the code that > there is work needed. > > Can we agree on that? It's the first time this has been requested, but it's your train-set and I will do whatever you ask. -- Lee Jones [李琼斯] Senior Technical Lead - Developer Services Linaro.org │ Open source software for Arm SoCs Follow Linaro: Facebook | Twitter | Blog
