Re: [PATCH] gpiolib: return -ENOENT when no GPIO mapping exists

From: Andy Shevchenko
Date: Mon Dec 09 2013 - 05:29:45 EST


On Fri, 2013-12-06 at 11:06 +0900, Alexandre Courbot wrote:
> Some devices drivers make use of optional GPIO parameters. For such
> drivers, it is important to discriminate between the case where no
> GPIO mapping has been defined for the function they are requesting, and
> the case where a mapping exists but an error occured while resolving it
> or when acquiring the GPIO.
>
> This patch changes the family of gpiod_get() functions such that they
> will return -ENOENT if and only if no GPIO mapping is defined for the
> requested function. Other error codes are used when an actual error
> occured during the GPIO resolution.
>

I like the idea.
One minor comment below (in code).

> Signed-off-by: Alexandre Courbot <acourbot@xxxxxxxxxx>
> ---
> I think this change should be merged early as not having it may prevent
> some users to switch to gpiod. I stumbled upon this issue while
> considering porting a simple driver (pwm_bl) that has an optional GPIO
> parameter.
>
> Mika, Andy: if Linus agrees with this change, could you take care of
> having -ENOENT returned as well for the ACPI and SFI GPIOs lookup?

I have already switched to -ENOENT, so, consider done.

> My understanding of ACPI was not sufficient to allow me to do it myself.
> SFI OTOH should be trivial as it is a simple table.
>
> Documentation/gpio/consumer.txt | 6 +++++-
> drivers/gpio/gpiolib.c | 31 ++++++++++++++++---------------
> 2 files changed, 21 insertions(+), 16 deletions(-)
>
> diff --git a/Documentation/gpio/consumer.txt b/Documentation/gpio/consumer.txt
> index 07c74a3765a0..e42f77d8d4ca 100644
> --- a/Documentation/gpio/consumer.txt
> +++ b/Documentation/gpio/consumer.txt
> @@ -38,7 +38,11 @@ device that displays digits), an additional index argument can be specified:
> const char *con_id, unsigned int idx)
>
> Both functions return either a valid GPIO descriptor, or an error code checkable
> -with IS_ERR(). They will never return a NULL pointer.
> +with IS_ERR() (they will never return a NULL pointer). -ENOENT will be returned
> +if and only if no GPIO has been assigned to the device/function/index triplet,
> +other error codes are used for cases where a GPIO has been assigned but an error
> +occured while trying to acquire it. This is useful to discriminate between mere
> +errors and an absence of GPIO for optional GPIO parameters.
>
> Device-managed variants of these functions are also defined:
>
> diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
> index 5fad38fcd701..e96d4a90c0c3 100644
> --- a/drivers/gpio/gpiolib.c
> +++ b/drivers/gpio/gpiolib.c
> @@ -2358,7 +2358,7 @@ static struct gpio_desc *gpiod_find(struct device *dev, const char *con_id,
> unsigned int idx,
> enum gpio_lookup_flags *flags)
> {
> - struct gpio_desc *desc = ERR_PTR(-ENODEV);
> + struct gpio_desc *desc = ERR_PTR(-ENOENT);
> struct gpiod_lookup_table *table;
> struct gpiod_lookup *p;
>
> @@ -2380,19 +2380,21 @@ static struct gpio_desc *gpiod_find(struct device *dev, const char *con_id,
> chip = find_chip_by_name(p->chip_label);
>
> if (!chip) {
> - dev_warn(dev, "cannot find GPIO chip %s\n",
> - p->chip_label);
> - continue;
> + dev_err(dev, "cannot find GPIO chip %s\n",
> + p->chip_label);
> + return ERR_PTR(-ENODEV);
> }
>
> if (chip->ngpio <= p->chip_hwnum) {
> - dev_warn(dev, "GPIO chip %s has %d GPIOs\n",
> - chip->label, chip->ngpio);
> - continue;
> + dev_err(dev, "requested GPIO %d but chip %s has %d\n",

The proposed message may confuse user. This lead to question in my head:
"what gpio chip has that referred by %d at the end of line".

Maybe something like "requested GPIO %d is out of range [0..%d] for chip
%s\n" ?

> + idx, chip->label, chip->ngpio);
> + return ERR_PTR(-EINVAL);
> }
>
> desc = gpiochip_offset_to_desc(chip, p->chip_hwnum);
> *flags = p->flags;
> +
> + return desc;
> }
>
> return desc;
> @@ -2404,7 +2406,8 @@ static struct gpio_desc *gpiod_find(struct device *dev, const char *con_id,
> * @con_id: function within the GPIO consumer
> *
> * Return the GPIO descriptor corresponding to the function con_id of device
> - * dev, or an IS_ERR() condition if an error occured.
> + * dev, -ENOENT if no GPIO has been assigned to the requested function, or
> + * another IS_ERR() code if an error occured while trying to acquire the GPIO.
> */
> struct gpio_desc *__must_check gpiod_get(struct device *dev, const char *con_id)
> {
> @@ -2421,7 +2424,9 @@ EXPORT_SYMBOL_GPL(gpiod_get);
> * This variant of gpiod_get() allows to access GPIOs other than the first
> * defined one for functions that define several GPIOs.
> *
> - * Return a valid GPIO descriptor, or an IS_ERR() condition in case of error.
> + * Return a valid GPIO descriptor, -ENOENT if no GPIO has been assigned to the
> + * requested function and/or index, or another IS_ERR() code if an error
> + * occured while trying to acquire the GPIO.
> */
> struct gpio_desc *__must_check gpiod_get_index(struct device *dev,
> const char *con_id,
> @@ -2446,13 +2451,9 @@ struct gpio_desc *__must_check gpiod_get_index(struct device *dev,
> * Either we are not using DT or ACPI, or their lookup did not return
> * a result. In that case, use platform lookup as a fallback.
> */
> - if (!desc || IS_ERR(desc)) {
> - struct gpio_desc *pdesc;
> + if (!desc || desc == ERR_PTR(-ENOENT)) {
> dev_dbg(dev, "using lookup tables for GPIO lookup");
> - pdesc = gpiod_find(dev, con_id, idx, &flags);
> - /* If used as fallback, do not replace the previous error */
> - if (!IS_ERR(pdesc) || !desc)
> - desc = pdesc;
> + desc = gpiod_find(dev, con_id, idx, &flags);
> }
>
> if (IS_ERR(desc)) {

--
Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
Intel Finland Oy

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/