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: <20230808145601.9401-1-larry.lai@yunjingtech.com> <20230808145601.9401-3-larry.lai@yunjingtech.com>
In-Reply-To: <20230808145601.9401-3-larry.lai@yunjingtech.com>
From:   Linus Walleij <linus.walleij@linaro.org>
Date:   Thu, 10 Aug 2023 13:41:29 +0200
Message-ID: <CACRpkdZZ020mQ8MuKzjab3_qwEcPuQ8=OcrRAz6OtN+B9_JgUw@mail.gmail.com>
Subject: Re: [PATCH V5 2/3] pinctrl: Add support pin control for UP board CPLD/FPGA
To:     "larry.lai" <larry.lai@yunjingtech.com>
Cc:     lee@kernel.org, andriy.shevchenko@linux.intel.com, pavel@ucw.cz,
        linux-kernel@vger.kernel.org, linux-gpio@vger.kernel.org,
        linux-leds@vger.kernel.org, GaryWang@aaeon.com.tw,
        musa.lin@yunjingtech.com, michael.wang@yunjingtech.com,
        jack.chang@yunjingtech.com, noah.hung@yunjingtech.com
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Hi Larry,

thanks for your patch!

I would really like Andy to look at this, because ACPI and I'm not good at
ACPI.

On Tue, Aug 8, 2023 at 4:57=E2=80=AFPM larry.lai <larry.lai@yunjingtech.com=
> wrote:

> The UP Squared board <http://www.upboard.com> implements certain
> features (pin control) through an on-board FPGA.
>
> Reported-by: kernel test robot <lkp@intel.com>
> Signed-off-by: Gary Wang <garywang@aaeon.com.tw>
> Signed-off-by: larry.lai <larry.lai@yunjingtech.com>
(...)

> +               if (strstr(pctrl->pctldesc->pins[offset[i]].name, "I2C") =
||
> +                   strstr(pctrl->pctldesc->pins[offset[i]].name, "PINMUX=
")) {

I think there is something like match a string list in the kernel,
Andy will know.

> +static int upboard_gpio_get(struct gpio_chip *gc, unsigned int offset)
> +{
> +       struct upboard_pinctrl *pctrl =3D container_of(gc, struct upboard=
_pinctrl, chip);
> +       unsigned int pin =3D pctrl->rpi_mapping[offset];
> +       int gpio =3D upboard_rpi_to_native_gpio(gc, offset);
> +
> +       if (gpio < 0)
> +               return gpio;
> +
> +       /* APL03 board open drain GPIO */
> +       if (pctrl->ident =3D=3D BOARD_UP_APL03) {
> +               int val =3D 0;
> +
> +               switch (pin) {
> +               case 0:
> +               case 1:
> +               case 9:
> +               case 23:
> +                       val =3D readl(pctrl->pins[pin].regs);
> +                       return val & 0x00000003;
> +               default:
> +                       break;
> +               }
> +       }
> +
> +       return gpiod_get_value(gpio_to_desc(gpio));

I don't really like this construction bouncing us back into the global GPIO
numberspace to get to the descriptor of the corresponding native GPIO.
gpio_to_desc() is an uglyhack and shall not be used.

You need to come up with something better here: request all the underlying
GPIO descriptors with gpiod_get(dev, .. ASIS) and store them in an array.
Look in e.g.
drivers/gpio/gpio-aggregator.c
drivers/gpio/gpio-latch.c
for examples of how we do this in other drivers.

The underlying GPIO lines need to be specified somehow, in worst case
using descriptor tables.

> +/* DMI Matches for older bios without fpga initialization */
> +static const struct dmi_system_id upboard_dmi_table[] __initconst =3D {
> +       {
> +               .ident =3D BOARD_U

A DMI table OK....

> +static struct platform_driver upboard_pinctrl_driver =3D {
> +       .driver =3D {
> +               .name =3D "upboard-pinctrl",
> +       },
> +};
> +module_platform_driver_probe(upboard_pinctrl_driver, upboard_pinctrl_pro=
be);

OK I don't know ACPI, but what I've seen the DMI table is supposed
to be used for some kind of lookup, such as checking in .init if the
device is there, so I don't get this. Also doesn't ACPI drivers use
the ACPI bus?

Yours,
Linus Walleij