Re: [PATCH] Documentation: kbuild: explain handling optional dependencies

From: Javier Martinez Canillas
Date: Wed Sep 13 2023 - 23:57:15 EST


"Arnd Bergmann" <arnd@xxxxxxxx> writes:

Hello Nicolas,

> On Wed, Sep 13, 2023, at 22:34, Nicolas Schier wrote:
>> On Wed, Sep 13, 2023 at 09:55:36PM +0200 Arnd Bergmann wrote:
>

[...]

>> I can choose between:
>>
>> BAR=y => FOO={N/m/y}
>> BAR=m => FOO is not selectable
>> BAR=n => FOO={N/m/y}
>
> That is indeed the point: if BAR=m, we want to be able to pick FOO=m
> here, otherwise it is impossible to enabled everything as modules.
>
> Another correct way to express the same thing as the first would
> be
>
> config FOO
> tristate "Support for foo hardware"
> depends on !BAR=m || m
>
> which I find even more confusing than the 'BAR || !BAR'
> convention, though we have that in a couple of places.
>
> I just found another variant that I had not seen before:
>
>> (Re-checked with BAR=IPV6 and FOO=WIREGUARD; CONFIG_WIREGUARD as 'depends on
>> IPV6 || !IPV6' in its kconfig definition, and both are tristate kconfig
>> symbols.)
>>

Which is correct because WIREGUARD can be built with IPV6 disabled, but
if both options are enabled then WIREGUARD can only be built-in if the
IPV6 option is also built-in.

WIREGUARD must be a module if IPV6 is also a module, but can still be a
module if IPV6 is built-in.

In other words, what this idiom express is that the following configs
are possible:

IPV6=n => WIREGUARD=y
IPV6=n => WIREGUARD=m
IPV6=y => WIREGUARD=y
IPV6=y => WIREGUARD=m
IPV6=m => WIREGUARD=m

but the following option is not possible:

IPV6=m => WIREGUARD=y

--
Best regards,

Javier Martinez Canillas
Core Platforms
Red Hat