Hi,

On 07/05/15 10:10, Paul Bolle wrote:

On Wed, 2015-05-06 at 11:31 +0200, Jens Kuske wrote:

quoted

--- a/drivers/pinctrl/sunxi/Kconfig
+++ b/drivers/pinctrl/sunxi/Kconfig

quoted

+config PINCTRL_SUN8I_H3
+	def_bool MACH_SUN8I
+	select PINCTRL_SUNXI_COMMON

quoted

--- a/drivers/pinctrl/sunxi/Makefile
+++ b/drivers/pinctrl/sunxi/Makefile

quoted

+obj-$(CONFIG_PINCTRL_SUN8I_H3)		+= pinctrl-sun8i-h3.o

PINCTRL_SUN8I_H3 is a bool symbol, so pinctrl-sun8i-h3.o will never be
part of a module, right?

(Note that PINCTRL_SUN8I_H3 appears to be an alias for MACH_SUN8I. Ie,
these two symbols operate in lockstep.)

quoted

--- /dev/null
+++ b/drivers/pinctrl/sunxi/pinctrl-sun8i-h3.c

quoted

+#include <linux/module.h>

quoted

+MODULE_DEVICE_TABLE(of, sun8i_h3_pinctrl_match);

quoted

+MODULE_AUTHOR("Jens Kuske [off-list ref]");
+MODULE_DESCRIPTION("Allwinner H3 pinctrl driver");
+MODULE_LICENSE("GPL");

This adds some module specific boilerplate. Was it perhaps your
intention to make PINCTRL_SUN8I_H3 a tristate symbol?

I don't know to be honest, I just followed the pattern of all the other
pinctrl-sun?i-*.c files. But it sounds logical that this is needless
in the current state.

Looks like it got introduced when splitting up the driver:
http://lists.infradead.org/pipermail/linux-arm-kernel/2014-April/251712.html

Maybe there were plans to use separate modules some day. If not, we
should remove it from the other files as well I guess. Maxime, could
you please comment on that?

Regards,
Jens