Received-SPF: pass (google.com: domain of linux-kernel+bounces-142524-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Precedence: bulk
MIME-Version: 1.0
References: <ZhgDCKhcHdwGoJ4Y@Z926fQmE5jqhFMgp6>
In-Reply-To: <ZhgDCKhcHdwGoJ4Y@Z926fQmE5jqhFMgp6>
From: Linus Walleij <linus.walleij@linaro.org>
Date: Fri, 12 Apr 2024 12:44:34 +0200
Message-ID: <CACRpkdbTdGajQLSGDkD0cWRG+79tpojYkdyF3t0jD7_WEDtQLw@mail.gmail.com>
Subject: Re: [PATCH][RFC][resend after bogus] gpio-syscon: do not report bogus error
To: linus.walleij@linaro.org, brgl@bgdev.pl, linux-gpio@vger.kernel.org, 
	linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

Hi Etienne,

thanks for your patch!

On Thu, Apr 11, 2024 at 5:35=E2=80=AFPM Etienne Buira <etienne.buira@free.f=
r> wrote:

> Do not call dev_err when gpio,syscon-dev is not set albeit unneeded.
> gpio-syscon is used with rk3328 chip, but this iomem region is
> documented in
> Documentation/devicetree/bindings/gpio/rockchip,rk3328-grf-gpio.yaml and
> does not look like to require gpio,syscon-dev setting.
>
> Signed-off-by: Etienne Buira <etienne.buira@free.fr>
> X-Prefers: kind explanations over rotten tomatoes

If you look in drivers/gpio/gpio-syscon.c you see this:

        priv->syscon =3D syscon_regmap_lookup_by_phandle(np, "gpio,syscon-d=
ev");
        if (IS_ERR(priv->syscon) && np->parent)
                priv->syscon =3D syscon_node_to_regmap(np->parent);

So the driver will attempt to grab the syscon from the parent if
it can't be located from a gpio,syscon-dev node.

But it's not optional, look in arch/arm64/boot/dts/rockchip/rk3328.dtsi:

        grf: syscon@ff100000 {
                compatible =3D "rockchip,rk3328-grf", "syscon", "simple-mfd=
";
                reg =3D <0x0 0xff100000 0x0 0x1000>;
(...)
                grf_gpio: gpio {
                        compatible =3D "rockchip,rk3328-grf-gpio";
                        gpio-controller;
                        #gpio-cells =3D <2>;
                };

So indeed the parent is a sycon, and syscon_node_to_regmap(np->parent) will
be used to populate priv->syscon on RK3328.

So what you could do insteaf of the kludge is something like:

bool has_parent_syscon =3D false;

priv->syscon =3D syscon_regmap_lookup_by_phandle(np, "gpio,syscon-dev");
if (IS_ERR(priv->syscon) && np->parent) {
        priv->syscon =3D syscon_regmap_lookup_by_phandle(np, "gpio,syscon-d=
ev");
        has_parent_syscon =3D true;
}
if (IS_ERR(priv->syscon))
        return PTR_ERR(priv->syscon);

Then when you get to the code you disable for the flag instead of:

if (!(priv->data->flags & GPIO_SYSCON_FEAT_NODEV)) {
(...)

instead do:

if (!has_parent_syscon) {
(...)

What do you think about this?

Yours,
Linus Walleij