On 05/09/2019 02.07, Miguel Ojeda wrote: > On Thu, Sep 5, 2019 at 2:00 AM Nick Desaulniers <[email protected]> > wrote: >> >> While you're here, would you mind replacing `__attribute__((unused))` >> with `__unused`? I would consider "naked attributes" (haven't been >> feature tested in include/linux/compiler_attributes.h and are verbose) >> to be an antipattern. > > +1 We should aim to avoid them entirely where possible. > > We have __always_unused and __maybe_unused, please choose whatever > fits best (both map to "unused", we don't have __unused).
Well, I agree in principle, but was trying to keep this minimal. FTR, if anything, I think the __attribute__((unused)) should simply be removed since it's implied by our (re)definition of inline/__inline/__inline__. Rasmus
