Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20220928064934.70867-1-renzhijie2@huawei.com> <YzPxMvjdjVEK1/b1@linutronix.de>
 <CAKXUXMxGt9UGhw9Ap_M3U2AF1vw2dX7WpDO71=UwV0Be3t4sNw@mail.gmail.com>
 <YzQQfW3h70OIiT14@linutronix.de> <e33cc08b-612b-4786-9b68-262c43af5ccb@www.fastmail.com>
In-Reply-To: <e33cc08b-612b-4786-9b68-262c43af5ccb@www.fastmail.com>
From:   Lukas Bulwahn <lukas.bulwahn@gmail.com>
Date:   Wed, 28 Sep 2022 13:43:51 +0200
Message-ID: <CAKXUXMxfF=WrVAWEZ=G8We+fOUvaM7o9uC_c8OVEDfa9Yv4sVA@mail.gmail.com>
Subject: Re: [PATCH -next] init/Kconfig: fix unmet direct dependencies
To:     Arnd Bergmann <arnd@arndb.de>
Cc:     Sebastian Andrzej Siewior <bigeasy@linutronix.de>,
        Ren Zhijie <renzhijie2@huawei.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        Nick Desaulniers <ndesaulniers@google.com>,
        Nathan Chancellor <nathan@kernel.org>,
        Vlastimil Babka <vbabka@suse.cz>,
        Masahiro Yamada <masahiroy@kernel.org>, seanjc@google.com,
        Johannes Weiner <hannes@cmpxchg.org>, ojeda@kernel.org,
        Masami Hiramatsu <mhiramat@kernel.org>,
        Dmitry Torokhov <dmitry.torokhov@gmail.com>,
        atomlin@redhat.com, ddiss@suse.de,
        Christophe Leroy <christophe.leroy@csgroup.eu>,
        linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Wed, Sep 28, 2022 at 11:32 AM Arnd Bergmann <arnd@arndb.de> wrote:
>
> On Wed, Sep 28, 2022, at 11:14 AM, Sebastian Andrzej Siewior wrote:
> > On 2022-09-28 09:20:42 [+0200], Lukas Bulwahn wrote:
> >> > Couldn't this become a depends?
> >> It could also be a depends (to resolve the warning).
> > =E2=80=A6
> >> It is just the question whether:
> >>
> >> When PROC_FS is not set, should the CHECKPOINT_RESTORE still be
> >> visible as a config option to add (and then automatically add
> >> PROC_FS)? Then select is right here.
> >
> > then CHECKPOINT_RESTORE is the only option selecting PROC_FS while
> > everyone else depends on it _or_ avoids using it in the absence of
> > PROC_FS.
>
> Right, we should not mix 'select' and 'depends on' for the same
> symbol, as that leads to circular dependencies and general
> confusion.
>
> If there is no way to use CHECKPOINT_RESTORE without procfs,
> then the symbol should just not be visible (it will still show
> up with the dependency when one searches in menuconfig).
> Force-enabling a major subsystem like procfs from another
> symbol is not a good solution.
>

Agree. I retract my Reviewed-by.

The arguments are clear to make this depend on PROC_FS.

Lukas