Re: [PATCH v5 2/2] gpio: add a reusable generic gpio_chip using regmap

From: Michael Walle
Date: Thu May 28 2020 - 09:00:41 EST


Am 2020-05-28 13:45, schrieb Andy Shevchenko:
On Thu, May 28, 2020 at 7:04 AM Michael Walle <michael@xxxxxxxx> wrote:

There are quite a lot simple GPIO controller which are using regmap to
access the hardware. This driver tries to be a base to unify existing
code into one place. This won't cover everything but it should be a good
starting point.

It does not implement its own irq_chip because there is already a
generic one for regmap based devices. Instead, the irq_chip will be
instantiated in the parent driver and its irq domain will be associate
to this driver.

For now it consists of the usual registers, like set (and an optional
clear) data register, an input register and direction registers.
Out-of-the-box, it supports consecutive register mappings and mappings
where the registers have gaps between them with a linear mapping between
GPIO offset and bit position. For weirder mappings the user can register
its own .xlate().

More comments from me below.

Thanks for the review.


...


# Device drivers. Generally keep list sorted alphabetically

Hmm...

+obj-$(CONFIG_GPIO_REGMAP) += gpio-regmap.o
obj-$(CONFIG_GPIO_GENERIC) += gpio-generic.o

...is it?

That's because gpio-regmap.o seems not be a driver and more of a generic
thing (like gpio-generic.o) and gpio-generic.o has another rule two lines
below and I don't want to put gpio-regmap.o in between.


...

+ * Copyright 2019 Michael Walle <michael@xxxxxxxx>

2020?

...

+#include <linux/gpio/driver.h>
+#include <linux/gpio-regmap.h>

Yes, I would like to see this as gpio/regmap.h (in gpio folder).

...

+static unsigned int gpio_regmap_addr(unsigned int addr)
+{
+ return (addr == GPIO_REGMAP_ADDR_ZERO) ? 0 : addr;

I would prefer rather to have
if (...)
return 0;

return addr;

here, but any of them fine.

yes looks nicer.


+}

...

+/**
+ * gpio_regmap_simple_xlate() - translate base/offset to reg/mask

Don't you get plenty of complains from kernel doc validation script?

now that I know there is one, yes. there are many complains.


You forgot to describe all function parameters here.

+ *
+ * Use a simple linear mapping to translate the offset to the bitmask.
+ */

This is a leftover, I'm actually gonna remove it since its not exported
anymore.


...

+ return (val & mask) ? 1 : 0;

Hmm... many (new!) GPIO drivers are using !! instead of ternary. Can
we do the same here?

ok


...

+static int gpio_regmap_get_direction(struct gpio_chip *chip,
+ unsigned int offset)
+{

+ if (gpio->reg_dir_out_base) {
+ base = gpio_regmap_addr(gpio->reg_dir_out_base);
+ invert = 0;
+ } else if (gpio->reg_dir_in_base) {
+ base = gpio_regmap_addr(gpio->reg_dir_in_base);
+ invert = 1;
+ } else {

+ return GPIO_LINE_DIRECTION_IN;

Hmm... Doesn't it an erroneous case and we basically shouldn't be here?

yeah, I'll return -EOPNOTSUPP. Better than just ignoring, right?


+ }

+ if (!!(val & mask) ^ invert)
+ return GPIO_LINE_DIRECTION_OUT;

+ else

Redundant 'else'.

IMHO, That looks really strange. like it has nothing to do with the
if statement. I'd like to keep that one.


+ return GPIO_LINE_DIRECTION_IN;
+}

+static int gpio_regmap_set_direction(struct gpio_chip *chip,
+ unsigned int offset, bool output)
+{

+ if (gpio->reg_dir_out_base) {
+ base = gpio_regmap_addr(gpio->reg_dir_out_base);
+ invert = 0;
+ } else if (gpio->reg_dir_in_base) {
+ base = gpio_regmap_addr(gpio->reg_dir_in_base);
+ invert = 1;
+ } else {

+ return 0;

Question as above.

same answer ;)


+ }

+ if (!invert)
+ val = (output) ? mask : 0;
+ else
+ val = (output) ? 0 : mask;

Why not positive conditional?
Also, too many parentheses.

ok


+ return regmap_update_bits(gpio->regmap, reg, mask, val);
+}

...

+/**
+ * gpio_regmap_register() - Register a generic regmap GPIO controller

+ *

Extra blank line.

didn't know that. so ok for all kind of these comments.


+ * @gpio: gpio_regmap device to register
+ *
+ * Returns 0 on success or an errno on failure.
+ */

...

+ if (!config->label)
+ chip->label = dev_name(config->parent);
+ else
+ chip->label = config->label;

Why not positive or here even ternary may look good

chip->label = config->label ?: dev_name(config->parent);

ok


...

+ ret = gpiochip_irqchip_add_domain(chip, config->irq_domain);

+ if (ret < 0)

Does ' < 0' make sense?

more or less, I'll change it to "if (ret)"


+ goto err_remove_gpiochip;

...

+/**
+ * gpio_regmap_unregister() - Unregister a generic regmap GPIO controller

+ *

Extra blank line

+ * @gpio: gpio_regmap device to unregister
+ */

...

+/**
+ * devm_gpio_regmap_register() - resource managed gpio_regmap_register()

+ *

Ditto.

+ * @dev: device that is registering this GPIO device
+ * @gpio: gpio_regmap device to register
+ *
+ * Managed gpio_regmap_register(). For generic regmap GPIO device registered by
+ * this function, gpio_regmap_unregister() is automatically called on driver
+ * detach. See gpio_regmap_register() for more information.
+ */

...

+ gpio = gpio_regmap_register(config);

+

Extra blank line.

ok

+ if (!IS_ERR(gpio)) {
+ *ptr = gpio;
+ devres_add(dev, ptr);
+ } else {
+ devres_free(ptr);
+ }

...

+#ifndef _LINUX_GPIO_REGMAP_H
+#define _LINUX_GPIO_REGMAP_H
+

Missed a lot, i.e.

struct device;
struct irq_domain;
struct regmap;

oops. right.


+struct gpio_regmap;
+
+#define GPIO_REGMAP_ADDR_ZERO ((unsigned long)(-1))
+#define GPIO_REGMAP_ADDR(addr) ((addr) ? : GPIO_REGMAP_ADDR_ZERO)
+
+/**
+ * struct gpio_regmap_config - Description of a generic regmap gpio_chip.

+ *

Extra blank line.

+ * @parent: The parent device
+ * @regmap: The regmap used to access the registers
+ * given, the name of the device is used
+ * @label: (Optional) Descriptive name for GPIO controller.
+ * If not given, the name of the device is used.
+ * @ngpio: Number of GPIOs

+ * @names: (Optional) Array of names for gpios

I don't see it. You really need to enable kernel doc validation warnings.

I've already noticed that in another mail.


+ * @reg_dat_base: (Optional) (in) register base address
+ * @reg_set_base: (Optional) set register base address
+ * @reg_clr_base: (Optional) clear register base address
+ * @reg_dir_in_base: (Optional) in setting register base address
+ * @reg_dir_out_base: (Optional) out setting register base address
+ * @reg_stride: (Optional) May be set if the registers (of the
+ * same type, dat, set, etc) are not consecutive.
+ * @ngpio_per_reg: Number of GPIOs per register
+ * @irq_domain: (Optional) IRQ domain if the controller is
+ * interrupt-capable
+ * @reg_mask_xlate: (Optional) Translates base address and GPIO
+ * offset to a register/bitmask pair. If not
+ * given the default gpio_regmap_simple_xlate()
+ * is used.
+ *
+ * The reg_mask_xlate translates a given base address and GPIO offset to

'reg_mask_xlate' -> '->reg_mask_xlate()' or '@reg_mask_xlate' or
special C function reference for kernel doc.

+ * register and mask pair. The base address is one of the given reg_*_base.

'reg_*_base' -> 'register base addresses in this structure' ?

+ *
+ * All base addresses may have the special value GPIO_REGMAP_ADDR_ZERO
+ * which forces the address to the value 0.

Also, since we have no separate documentation, describe the rules,
that some of the registers are actually required, and some maybe in
conflict (these rules you have in register function).

ok

-michael