Re: [PATCH v3 4/6] binder: Scaffolding for binder_alloc KUnit tests

[Kees Cook <kees@kernel.org>]

On Wed, Jul 16, 2025 at 10:42:58PM +0000, Carlos Llamas wrote:
> On Wed, Jul 16, 2025 at 03:28:49PM -0700, Tiffany Yang wrote:
> > Kees Cook <kees@kernel.org> writes:
> > 
> > > > ...
> > 
> > > I'm used to the "#ifdef CONFIG_..." idiom, but looking at the tree, I
> > > see that "#if IS_ENANLED(CONFIG...)" is relatively common too. I don't
> > > think there is a function difference, so I leave the style choice up to
> > > you! ;)
> > 
> > 
> > IIRC, I had tried using the #ifdef to inject test-specific code in some
> > places, but that created issues when we were loading KUnit as a module
> > because it causes the built-in code to be built as though the
> > "CONFIG_..." is undefined. Consequently, I started using IS_ENABLED by
> > default, but I'm not sure if it's strictly necessary for exposing
> > functions with a header file since I'd assume "CONFIG_..." will be
> > defined when we're building the module?
> 
> Right, IS_ENABLED() is the "short-hand" for ...
> 	#if defined(CONFIG_x) || defined(CONFIG_x_MODULE)
> ... which is what we need here.

Ah-ha! Okay, thanks for the explanation. I don't do much module work,
it seems. ;)

-- 
Kees Cook