Re: [PATCH 2/11] x86: convert to generic helpers for IPI function calls

[Jens Axboe <jens.axboe@oracle.com>]

On Tue, Apr 22 2008, Jens Axboe wrote:
> On Tue, Apr 22 2008, Sam Ravnborg wrote:
> > On Tue, Apr 22, 2008 at 09:57:28AM +0200, Jens Axboe wrote:
> > > This converts x86 and x86-64 to use the new helpers for
> > > smp_call_function() and friends, and adds support for
> > > smp_call_function_single().
> > > 
> > > diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
> > > index 87a693c..aab97e3 100644
> > > --- a/arch/x86/Kconfig
> > > +++ b/arch/x86/Kconfig
> > > @@ -156,6 +156,11 @@ config GENERIC_PENDING_IRQ
> > >  	depends on GENERIC_HARDIRQS && SMP
> > >  	default y
> > >  
> > > +config GENERIC_SMP_HELPERS
> > > +	bool
> > > +	depends on SMP
> > > +	default y
> > > +
> > 
> > Hi Jens.
> > 
> > [Not x86 specific - it was the first patch
> > touching a Kconfig file]
> > 
> > Could we define a single config variable
> > in for example kernel/Kconfig(*)
> > like this:
> > 
> > config HAVE_GENERIC_SMP_HELPERS
> > 	def_bool n
> > 
> > And then the archs that uses the generic SMP helpers just
> > do a simple:
> > 
> >  config X86
> > +	select HAVE_GENERIC_SMP_HELPERS
> > 
> > 
> > This is the recommended method today and is documented
> > in Documentation/kbuild/kconfig-language.txt if
> > you need a bit more intro.
> > 
> > (*) I know we do not have kernel/Kconfig today.
> >     But maybe this is a good time to add it and
> >     source it from arch/Kconfig
> 
> Ah yes, that looks like a better approach. I'll adapt the patches to
> this method instead of selecting smp.o, thanks.

I added

config USE_GENERIC_SMP_HELPERS
        def_bool n

in arch/Kconfig and added a 'select USE_GENERIC_SMP_HELPERS' in the arch
private Kconfig files instead. It also reduces the line count further,
thanks for the suggestion! It's cleaner this way.

-- 
Jens Axboe