Re: [PULL] Networking for v6.3

From: Alexander Lobakin <aleksander.lobakin@intel.com>
To: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Jakub Kicinski <kuba@kernel.org>,
	Arkadiusz Kubalewski <arkadiusz.kubalewski@intel.com>,
	<davem@davemloft.net>, <netdev@vger.kernel.org>,
	<linux-kernel@vger.kernel.org>, <pabeni@redhat.com>,
	<bpf@vger.kernel.org>, <ast@kernel.org>
Subject: Re: [PULL] Networking for v6.3
Date: Wed, 22 Feb 2023 20:07:21 +0100	[thread overview]
Message-ID: <d04f822b-ac63-7f45-ef4d-978876d57307@intel.com> (raw)
In-Reply-To: <CAHk-=wi_410KZqHwF-WL5U7QYxnpHHHNP-3xL=g_y89XnKc-uw@mail.gmail.com>

From: Linus Torvalds <torvalds@linux-foundation.org>
Date: Tue, 21 Feb 2023 18:46:26 -0800

> On Tue, Feb 21, 2023 at 3:38 PM Jakub Kicinski <kuba@kernel.org> wrote:
>>
>>   git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git tags/net-next-6.3
> 
> Ok, so this is a bit nitpicky, but commit c7ef8221ca7d ("ice: use GNSS
> subsystem instead of TTY") ends up doing odd things to kernel configs.
> 
> My local configuration suddenly grew this:
> 
>     CONFIG_ICE_GNSS=y
> 
> which is pretty much nonsensical.
> 
> The reason? It's defined as
> 
>     config ICE_GNSS
>             def_bool GNSS = y || GNSS = ICE
> 
> and so it gets set even when both GNSS and ICE are both disabled,
> because 'n' = 'n'.

It was me who originally suggested this and I forgot ice doesn't have
its own Kconfig file and doesn't guard its options under `if ICE`, sorry
=\ It would work if it had.

> 
> Does it end up *mattering*? No. It's only used in the ICE driver, but
> it really looks all kinds of odd, and it makes the resulting .config
> files illogical.
> 
> Maybe I'm the only one who looks at those things. I do it because I

FWIW I do as well, ironically enough :D
But I have `CONFIG_ICE=m` on all of my systems, so there was no chance
for me to notice this. Thanks!

> think they are sometimes easier to just edit directly, but also
> because for me it's a quick way to see if somebody has sneaked in new
> config options that are on by default when they shouldn't be.
> 
> I'd really prefer to not have the resulting config files polluted with
> nonsensical config options.
> 
> I suspect it would be as simple as adding a
> 
>         depends on ICE != n
> 
> to that thing, but I didn't get around to testing that. I thought it
> would be better to notify the guilty parties.

Patch is on its way already, it just drops the opt and uses CONFIG_GNSS
directly.

> 
> Anyway, this has obviously not held up me pulling the networking
> changes, and you should just see this as (yet another) sign of "yeah,
> Linus cares about those config files to a somewhat unhealthy degree".
> 
>                       Linus
> 

Olek