Re: [PATCH v2 5/8] pinctrl: Add core support for Aspeed SoCs

From: Linus Walleij
Date: Mon Aug 22 2016 - 09:45:55 EST


On Fri, Aug 19, 2016 at 2:44 PM, Andrew Jeffery <andrew@xxxxxxxx> wrote:

> +++ b/drivers/pinctrl/aspeed/Kconfig
> @@ -0,0 +1,8 @@
> +config PINCTRL_ASPEED
> + bool
> + depends on (ARCH_ASPEED || COMPILE_TEST) && OF
> + select PINMUX
> + select PINCONF
> + select GENERIC_PINCONF
> + select MFD_SYSCON
> + select REGMAP_MMIO

Since this device is spawn from the syscon, should it not be
"depends on MFD_SYSCON"?

(No big deal, if you think this is right then go with it.)

> +#include <linux/io.h>

What is this include for?

> +#include <linux/mfd/syscon.h>
> +#include <linux/platform_device.h>
> +#include <linux/slab.h>
> +#include <linux/string.h>
> +#include "../core.h"
> +#include "pinctrl-aspeed.h"

No #include <linux/regmap.h>?

Maybe some #includes are centralized to pinctrl-aspeed.h
I don't know, just make sure you don't have implicit includes.

> + if (regmap_read(map, desc->reg, &val) < 0)
> + return false;
> +
> + val &= ~desc->mask;
> + val |= pattern << __ffs(desc->mask);
> +
> + if (regmap_write(map, desc->reg, val) < 0)
> + return false;

Use:

regmap_update_bits(map,
desc->reg,
desc->mask,
pattern << __ffs->desc->mask);

Or something like that instead of reimplementing
mask-and-set. Regmap already knows how to do the
business.

(Applied everywhere in the driver where you have a
mask-and-set like this).

The expression core engine is still a complete mystery
for me, I will just trust you that it works as intended.

Yours,
Linus Walleij