Re: [PATCH v1 1/1] RAS/AMD/ATL: Drop malformed default N from Kconfig

[Andy Shevchenko]

On Wed, May 13, 2026 at 07:02:40PM +0200, Borislav Petkov wrote:
> On Wed, May 13, 2026 at 07:31:24PM +0300, Andy Shevchenko wrote:
> > On Wed, May 13, 2026 at 06:27:44PM +0200, Borislav Petkov wrote:
> > > On Wed, May 13, 2026 at 06:22:53PM +0200, Andy Shevchenko wrote:
> > > > First of all, it has to be 'default n' (small letter n), otherwise
> > > > it looks for CONFIG_N which is absent and in case of appearance
> > 
> > > Who does look for CONFIG_N?
> > 
> > Hmm... I need to rephrase that. Any suggestion how to refer to the Kconfig
> > symbol in a clearer way?
> 
> So I tried searching for CONFIG_N in "make menuconfig" and it gave me many
> results, but with or without your patch I didn't see a difference.

And it has to be no difference (with all my hopes we do *not* have 'config N').

> Just point to the "default" grammar definition - I see
> 
>     config NEW_NAME
>         bool "New option name"
>         default OLD_NAME
> 
> in Documentation/kbuild/kconfig-language.rst which says that default <symbols
> in caps> refers to a symbol but the grammar should be properly documented
> somewhere too (can't find it now...).

Yeah, but here is the bug report now for dangling Kconfig symbols
https://bugzilla.kernel.org/show_bug.cgi?id=216748
this patch addresses one of the complains the bug report has.

-- 
With Best Regards,
Andy Shevchenko