def_bool n

def_bool n

[Jan Beulich]

Could either of you clarify what the significance of such a construct
(encountered several times in arch/x86/Kconfig alone) is?

At first, I noticed it only with GENERIC_LOCKBREAK (it is my
understanding that this block can go away entirely), but then realized
that with the bool->def_bool conversion this was introduced in various
other places (where "default n" was used before, which seems as
pointless a statement). Am I missing something?

Thanks, Jan

Re: def_bool n

[Sam Ravnborg]

On Mon, Aug 04, 2008 at 01:29:06PM +0100, Jan Beulich wrote:
> Could either of you clarify what the significance of such a construct
> (encountered several times in arch/x86/Kconfig alone) is?
> 
> At first, I noticed it only with GENERIC_LOCKBREAK (it is my
> understanding that this block can go away entirely), but then realized
> that with the bool->def_bool conversion this was introduced in various
> other places (where "default n" was used before, which seems as
> pointless a statement). Am I missing something?

No.

arch/x86/Kconfig:
config GENERIC_LOCKBREAK
        def_bool n

I assume this is solely to document that x86 does
NOT use GENERIC_LOCKBREAK as it serves no other purpose.

Notice that we may see:

init/Kconfig:
config FOO
	def_bool y

arch/$ARCH/Kconfig:
config FOO
	def_bool n

And this will result in setting FOO to the last seen value.
All the HAVE_* stuff is btw to avoid such double definition
of a config symbol.

GENERIC_LOCKBREAK is another candidate.

	Sam