Re: [PATCH] mm: memcontrol.c: move mem_cgroup_id_get_many under CONFIG_MMU

[Chris Down]

Michal Hocko writes:
> yes, I would just ignore this warning. Btw. it seems that this is
> enabled by default for -Wall. Is this useful for kernel builds at
> all? Does it realistically help discovering real issues? If not then
> can we simply blacklist it?

There's no way we're the first people to encounter these problems, so what did 
we do in the past when situations like this (adding a generic API which is not 
yet used by non-configurable code) came up, and in retrospect did they work 
well?

As far as I know -Wunused-function also guards against other errors, like when 
a function is prototyped but not actually defined, which might be more useful 
to know about.

(Side note: I'm moderately baffled that a tightly scoped __maybe_unused is 
considered sinister but somehow disabling -Wunused-function is on the table 
:-))