Re: [PATCH] x86/Kconfig: Further simplify the NR_CPUS config

[Ingo Molnar]

* Randy Dunlap <rdunlap@xxxxxxxxxxxxx> wrote:

> On 02/10/2018 02:19 PM, Linus Torvalds wrote:
> > Looks good to me.
> > 
> > At the risk of bike-shedding, we could remove all the
> > 
> >         default    1 if !SMP
> > 
> > from the BEGIN/END/DEFAULT things, and perhaps just keep that part in NR_CPUS.
> > 
> > I didn't check, but I *think* it would work to just do
> > 
> >  config NR_CPUS
> >         int "Maximum number of CPUs" if SMP && !MAXSMP
> >         range NR_CPUS_RANGE_BEGIN NR_CPUS_RANGE_END
> >         default "1" if !SMP
> >         default NR_CPUS_DEFAULT
> > 
> > but maybe the "range" line would need an "if !SMP" on it too to avoid
> > the issue with "1" being out of range.,
> 
> Yeah, I had an early test that failed due to something like that.

I *think* I slightly prefer the current approach, because while it's somewhat 
verbose, the advantage is that this way we have *all* range considerations for a 
given main hardware variant in a single place, and the main NR_CPUS config entry 
is 'passive' in terms of determining the range used.

Plus the verbosity isn't really a problem either, as the whole approach is a 
'verbose' expansion of an overly complex config expression, for better 
readability/maintainability.

Thanks,

	Ingo