On Fri, 28 Sep 2018, Max Filippov wrote:
> On Fri, Sep 28, 2018 at 2:08 PM, Andy Lutomirski <luto@amacapital.net> wrote:
> >> On Sep 28, 2018, at 1:26 PM, Thomas Gleixner <tglx@linutronix.de> wrote:
> >>
> >>> On Fri, 28 Sep 2018, Max Filippov wrote:
> >>>
> >>>> On Fri, Sep 28, 2018 at 11:02 AM, Andy Lutomirski <luto@amacapital.net> wrote:
> >>>> There may be a much nicer solution.  Unless I missed something, only
> >>>> mips and xtensa even have the possibility of cmpxchg being missing.
> >>>> We could just make those arches supply a futex-detecting helper.
> >>>
> >>> In case of xtensa availability of cmpxchg is known at build time.
> >>
> >> That makes it even simpler. Could you provide a patch which selects
> >> CONFIG_HAVE_FUTEX_CMPXCHG for the right set of CPUs please?
> >>
> >>
> >
> > I think that’s the wrong approach, since it won’t cover mips. How about adding this to mips and xtensa only:
> >
> > static inline void arch_have_futex_cmpxchg(void) {...};
> > #define arch_have_futex_cmpxchg arch_have_futex_cmpxchg
> >
> > And getting rid of the config option.
> 
> I'd rather do that, given that defining Kconfig entries that describe
> parts of xtensa configuration is somewhat awkward and redundant.

Fair enough.

Thanks,

	tglx