* [PATCH] gpio: remove less important #ifdef around declarations
@ 2019-07-07 2:30 Masahiro Yamada
2019-07-28 22:30 ` Linus Walleij
0 siblings, 1 reply; 2+ messages in thread
From: Masahiro Yamada @ 2019-07-07 2:30 UTC (permalink / raw)
To: linux-gpio, Linus Walleij
Cc: Masahiro Yamada, Bartosz Golaszewski, linux-kernel
The whole struct/function declarations in this header are surrounded
by #ifdef.
As far as I understood, the motivation of doing so is probably to break
the build earlier if a driver misses to select or depend on correct
CONFIG options in Kconfig.
Since commit 94bed2a9c4ae ("Add -Werror-implicit-function-declaration")
no one cannot call functions that have not been declared.
So, I see some benefit in doing this in the cost of uglier headers.
In reality, it would not be so easy to catch missed 'select' or
'depends on' because GPIOLIB, GPIOLIB_IRQCHIP etc. are already selected
by someone else eventually. So, this kind of error, if any, will be
caught by randconfig bots.
In summary, I am not a big fan of cluttered #ifdef nesting, and this
does not matter for normal developers. The code readability wins.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
---
FYI,
If you want replace the commit log of the pinctrl variant,
the better log is available:
https://lkml.org/lkml/2019/6/26/1344
include/linux/gpio/driver.h | 24 +++++++-----------------
1 file changed, 7 insertions(+), 17 deletions(-)
diff --git a/include/linux/gpio/driver.h b/include/linux/gpio/driver.h
index a1d273c96016..845d657f2431 100644
--- a/include/linux/gpio/driver.h
+++ b/include/linux/gpio/driver.h
@@ -19,9 +19,6 @@ struct gpio_device;
struct module;
enum gpiod_flags;
-#ifdef CONFIG_GPIOLIB
-
-#ifdef CONFIG_GPIOLIB_IRQCHIP
/**
* struct gpio_irq_chip - GPIO interrupt controller
*/
@@ -167,7 +164,6 @@ struct gpio_irq_chip {
*/
void (*irq_disable)(struct irq_data *data);
};
-#endif
/**
* struct gpio_chip - abstract a GPIO controller
@@ -445,16 +441,12 @@ bool gpiochip_line_is_valid(const struct gpio_chip *chip, unsigned int offset);
/* get driver data */
void *gpiochip_get_data(struct gpio_chip *chip);
-struct gpio_chip *gpiod_to_chip(const struct gpio_desc *desc);
-
struct bgpio_pdata {
const char *label;
int base;
int ngpio;
};
-#if IS_ENABLED(CONFIG_GPIO_GENERIC)
-
int bgpio_init(struct gpio_chip *gc, struct device *dev,
unsigned long sz, void __iomem *dat, void __iomem *set,
void __iomem *clr, void __iomem *dirout, void __iomem *dirin,
@@ -467,10 +459,6 @@ int bgpio_init(struct gpio_chip *gc, struct device *dev,
#define BGPIOF_READ_OUTPUT_REG_SET BIT(4) /* reg_set stores output value */
#define BGPIOF_NO_OUTPUT BIT(5) /* only input */
-#endif
-
-#ifdef CONFIG_GPIOLIB_IRQCHIP
-
int gpiochip_irq_map(struct irq_domain *d, unsigned int irq,
irq_hw_number_t hwirq);
void gpiochip_irq_unmap(struct irq_domain *d, unsigned int irq);
@@ -559,15 +547,11 @@ static inline int gpiochip_irqchip_add_nested(struct gpio_chip *gpiochip,
}
#endif /* CONFIG_LOCKDEP */
-#endif /* CONFIG_GPIOLIB_IRQCHIP */
-
int gpiochip_generic_request(struct gpio_chip *chip, unsigned offset);
void gpiochip_generic_free(struct gpio_chip *chip, unsigned offset);
int gpiochip_generic_config(struct gpio_chip *chip, unsigned offset,
unsigned long config);
-#ifdef CONFIG_PINCTRL
-
/**
* struct gpio_pin_range - pin range controlled by a gpio chip
* @node: list for maintaining set of pin ranges, used internally
@@ -580,6 +564,8 @@ struct gpio_pin_range {
struct pinctrl_gpio_range range;
};
+#ifdef CONFIG_PINCTRL
+
int gpiochip_add_pin_range(struct gpio_chip *chip, const char *pinctl_name,
unsigned int gpio_offset, unsigned int pin_offset,
unsigned int npins);
@@ -620,6 +606,10 @@ void gpiochip_free_own_desc(struct gpio_desc *desc);
void devprop_gpiochip_set_names(struct gpio_chip *chip,
const struct fwnode_handle *fwnode);
+#ifdef CONFIG_GPIOLIB
+
+struct gpio_chip *gpiod_to_chip(const struct gpio_desc *desc);
+
#else /* CONFIG_GPIOLIB */
static inline struct gpio_chip *gpiod_to_chip(const struct gpio_desc *desc)
@@ -631,4 +621,4 @@ static inline struct gpio_chip *gpiod_to_chip(const struct gpio_desc *desc)
#endif /* CONFIG_GPIOLIB */
-#endif
+#endif /* __LINUX_GPIO_DRIVER_H */
--
2.17.1
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH] gpio: remove less important #ifdef around declarations
2019-07-07 2:30 [PATCH] gpio: remove less important #ifdef around declarations Masahiro Yamada
@ 2019-07-28 22:30 ` Linus Walleij
0 siblings, 0 replies; 2+ messages in thread
From: Linus Walleij @ 2019-07-28 22:30 UTC (permalink / raw)
To: Masahiro Yamada
Cc: open list:GPIO SUBSYSTEM, Bartosz Golaszewski, linux-kernel
On Sun, Jul 7, 2019 at 4:31 AM Masahiro Yamada
<yamada.masahiro@socionext.com> wrote:
> The whole struct/function declarations in this header are surrounded
> by #ifdef.
>
> As far as I understood, the motivation of doing so is probably to break
> the build earlier if a driver misses to select or depend on correct
> CONFIG options in Kconfig.
>
> Since commit 94bed2a9c4ae ("Add -Werror-implicit-function-declaration")
> no one cannot call functions that have not been declared.
>
> So, I see some benefit in doing this in the cost of uglier headers.
>
> In reality, it would not be so easy to catch missed 'select' or
> 'depends on' because GPIOLIB, GPIOLIB_IRQCHIP etc. are already selected
> by someone else eventually. So, this kind of error, if any, will be
> caught by randconfig bots.
>
> In summary, I am not a big fan of cluttered #ifdef nesting, and this
> does not matter for normal developers. The code readability wins.
>
> Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Hm I guess you're right.
This patch does not apply cleanly on v5.3-rc1, could you rebase it
and resend?
Yours,
Linus Walleij
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2019-07-28 22:30 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-07-07 2:30 [PATCH] gpio: remove less important #ifdef around declarations Masahiro Yamada
2019-07-28 22:30 ` Linus Walleij
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).