Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <cover.1589623456.git.baolin.wang7@gmail.com> <db5adf754300bdc89d561ea3fb23afc4b6a3ee87.1589623456.git.baolin.wang7@gmail.com>
 <CAK8P3a38RVe0=i2pObgaOn+HP8cpdKefiac6DcyXuKnyH4xOyA@mail.gmail.com>
In-Reply-To: <CAK8P3a38RVe0=i2pObgaOn+HP8cpdKefiac6DcyXuKnyH4xOyA@mail.gmail.com>
From:   Orson Zhai <orsonzhai@gmail.com>
Date:   Thu, 21 May 2020 09:08:35 +0800
Message-ID: <CA+H2tpFyCoHHVRM1wcn7TKBwxkafg6ZAVWzrsP3wZRNH5kPR7A@mail.gmail.com>
Subject: Re: [PATCH v4 1/2] mfd: syscon: Support physical regmap bus
To:     Arnd Bergmann <arnd@arndb.de>
Cc:     Baolin Wang <baolin.wang7@gmail.com>,
        Lee Jones <lee.jones@linaro.org>,
        Mark Brown <broonie@kernel.org>,
        Lyra Zhang <zhang.lyra@gmail.com>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, May 19, 2020 at 9:19 PM Arnd Bergmann <arnd@arndb.de> wrote:
>
> On Sat, May 16, 2020 at 12:13 PM Baolin Wang <baolin.wang7@gmail.com> wrote:
> >
> > Some platforms such as Spreadtrum platform, define a special method to
> > update bits of the registers instead of reading and writing, which means
> > we should use a physical regmap bus to define the reg_update_bits()
> > operation instead of the MMIO regmap bus.
> >
> > Thus add a a __weak function  for the syscon driver to allow to register
> > a physical regmap bus to support this new requirement.
> >
> >  };
> >
> > +struct regmap * __weak syscon_regmap_init(struct device_node *np,
> > +                                         void __iomem *base,
> > +                                         struct regmap_config *syscon_config)
> > +{
> > +       return regmap_init_mmio(NULL, base, syscon_config);
> > +}
> > +
>
> Sorry, I don't think the __weak function is going to help here. I'm not
> sure whether it actually does what you want when both syscon and
> sprd_syscon are loadable modules (I would guess not), but it clearly
> won't work when syscon is built-in and sprd_syscon is a module, and
> even if the module loader knows how to resolve __weak symbols,


I happened to see module.c last week. It seems there's some code to handle
weak symbols by checking about STB_WEAK. But it is not in this case.
When syscon.c is built-in and sprd_syscon is module, this would not
work because the
__weak one has been built into kernel already. Maybe live patch
feature may work it out.
But it seems over expensive  to do things like this.

> I would not want to rely on module load ordering to make it behave
> the right way.

Agree. All other sprd modules would must be loaded behind sprd_syscon
and this will not be
handled by modprobe.

Thanks to pointing out this.

Best,
Orson

>
>       Arnd