Hi Mickael

Subject: Re: [PATCH v4] gpio: add gpio-hog support Importance: High

On 12. 06. 19 6:11, Heiko Schocher wrote:

...

add gpio-hog support. GPIO hogging is a mechanism providing automatic GPIO request and configuration as part of the gpio-controller's driver probe function.

for more infos see: doc/device-tree-bindings/gpio/gpio.txt

Signed-off-by: Heiko Schocher hs@denx.de

---

clean travis build, see result: https://travis-ci.org/hsdenx/u-boot-test/builds/544040419

Changes in v4:

• rework as suggested by Patrick based on patch: http://patchwork.ozlabs.org/patch/1098913/ use new UCLASS_NOP hog detection + DT parsing in gpio_post_bind() .... info saved in platdata gpio configuration for hog (request and set_dir) => probe function

Changes in v3:

• probe now all gpio devices with gpio-hogs from board_r before board_late_init() call or if someone wants to access gpio-hog based gpios with gpio_hog_lookup_name() This is needed, because we cannot be sure, if a gpio device which contains gpio-hogs is probed.
• add line-name property as Michal recommended
• renamed gpio_priv_one into gpio_priv_one_hog and protected through CONFIG_DM_GPIO_HOG

Changes in v2:

• move gpio_hog() call from post_probe() to post_bind() call. Check if current gpio device has gpio-hog definitions, if so, probe it.

common/board_r.c | 6 + doc/device-tree-bindings/gpio/gpio.txt | 55 ++++++++ drivers/gpio/Kconfig | 10 ++ drivers/gpio/gpio-uclass.c | 181 ++++++++++++++++++++++--- include/asm-generic/gpio.h | 32 +++++ 5 files changed, 268 insertions(+), 16 deletions(-)

diff --git a/common/board_r.c b/common/board_r.c index 150e8cd424..507da4c74a 100644 --- a/common/board_r.c +++ b/common/board_r.c @@ -49,6 +49,9 @@ #include <linux/err.h> #include <efi_loader.h> #include <wdt.h> +#if defined(CONFIG_DM_GPIO_HOG) +#include <asm/gpio.h> +#endif

DECLARE_GLOBAL_DATA_PTR;

@@ -796,6 +799,9 @@ static init_fnc_t init_sequence_r[] = { #ifdef CONFIG_CMD_NET initr_ethaddr, #endif +#if defined(CONFIG_DM_GPIO_HOG)

• gpio_hog_probe_all,

+#endif

Any reason to have hog at this place ? Before board_late_init....

I expect more before board_init() for my future use-case, to replace the current (dirty) implementation with pinctrl-0

#if defined(CONFIG_WDT) initr_watchdog, #endif +#if defined(CONFIG_DM_GPIO_HOG) + gpio_hog_probe_all, +#endif #if defined(CONFIG_ARM) || defined(CONFIG_NDS32) || defined(CONFIG_RISCV) || \ defined(CONFIG_SANDBOX) board_init, /* Setup chipselects */ #endif

In my device tree I have :

stmfx: stmfx@42 { compatible = "st,stmfx-0300"; reg = <0x42>; interrupts = <8 IRQ_TYPE_EDGE_RISING>; interrupt-parent = <&gpioi>; vdd-supply = <&v3v3>;

stmfx_pinctrl: stmfx-pin-controller { compatible = "st,stmfx-0300-pinctrl"; gpio-controller; #gpio-cells = <2>; interrupt-controller; #interrupt-cells = <2>; gpio-ranges = <&stmfx_pinctrl 0 0 24>; pinctrl-names = "default"; pinctrl-0 = <&hog_pins>;

hog_pins: hog { pins = "gpio14"; drive-push-pull; bias-pull-down; }; };

And in board, I force probe to configure HOG....before to initialize the DISPLAY, so in

./board/st/stm32mp1/stm32mp1.c : board _init()

/* board dependent setup after realloc */ int board_init(void) { ..... /* probe all PINCTRL for hog */ for (uclass_first_device(UCLASS_PINCTRL, &dev); dev; uclass_next_device(&dev)) { pr_debug("probe pincontrol = %s\n", dev->name); }

Or you can let each board choose when call the function as it is already done for other function as regulators_enable_boot_on() So don't add it in init_sequence_r[] / board_r.c ?...

Anyway it is not blocking for me, it just a question.

...

#ifdef CONFIG_BOARD_LATE_INIT board_late_init, #endif diff --git a/doc/device-tree-bindings/gpio/gpio.txt b/doc/device-tree-bindings/gpio/gpio.txt index f7a158d858..e774439369 100644 --- a/doc/device-tree-bindings/gpio/gpio.txt +++ b/doc/device-tree-bindings/gpio/gpio.txt @@ -210,3 +210,58 @@ Example 2: Here, three GPIO ranges are defined wrt. two pin controllers. pinctrl1 GPIO ranges are defined using pin numbers whereas the GPIO ranges wrt. pinctrl2 are named "foo" and "bar".

+3) GPIO hog definitions +-----------------------

+The GPIO chip may contain GPIO hog definitions. GPIO hogging is a +mechanism providing automatic GPIO request and configuration as part +of the gpio-controller's driver probe function.

+Each GPIO hog definition is represented as a child node of the GPIO

controller.

...

+Required properties: +- gpio-hog: A property specifying that this child node represents a GPIO hog. +- gpios: Store the GPIO information (id, flags) for the GPIO to

•      affect.
  

•          ! Not yet support more than one gpio !
  

+Only one of the following properties scanned in the order shown below. +- input: A property specifying to set the GPIO direction as input. +- output-low A property specifying to set the GPIO direction as output with

•      the value low.
  

+- output-high A property specifying to set the GPIO direction as output with

•      the value high.
  

+Optional properties: +- line-name: The GPIO label name. If not present the node name is used.

+Example:

•    tca6416@20 {
  

•            compatible = "ti,tca6416";
  

•            reg = <0x20>;
  

•            #gpio-cells = <2>;
  

•            gpio-controller;
  

•            env_reset {
  

•                    gpio-hog;
  

•                    input;
  

•                    gpios = <6 GPIO_ACTIVE_LOW>;
  

•            };
  

•            boot_rescue {
  

•                    gpio-hog;
  

•                    input;
  

•                    gpios = <7 GPIO_ACTIVE_LOW>;
  

•            };
  

•    };
  

+For the above Example you can than access the gpio in your boardcode +with:

•    desc = gpio_hog_lookup_name("boot_rescue.gpio-hog");
  

•    if (desc) {
  

•            if (dm_gpio_get_value(desc))
  

•                    printf("\nBooting into Rescue System\n");
  

• else
  

• 	printf("\nBoot normal\n");
  

diff --git a/drivers/gpio/Kconfig b/drivers/gpio/Kconfig index e36a8abc42..fa1c99700f 100644 --- a/drivers/gpio/Kconfig +++ b/drivers/gpio/Kconfig @@ -14,6 +14,16 @@ config DM_GPIO particular GPIOs that they provide. The uclass interface is defined in include/asm-generic/gpio.h.

+config DM_GPIO_HOG

• bool "Enable GPIO hog support"
• depends on DM_GPIO
• default n
• help

•  Enable gpio hog support
  

•  The GPIO chip may contain GPIO hog definitions. GPIO hogging
  

•  is a mechanism providing automatic GPIO request and config-
  

•  uration as part of the gpio-controller's driver probe function.
  

config ALTERA_PIO bool "Altera PIO driver" depends on DM_GPIO diff --git a/drivers/gpio/gpio-uclass.c b/drivers/gpio/gpio-uclass.c index da5e9ba6e5..308d0863ad 100644 --- a/drivers/gpio/gpio-uclass.c +++ b/drivers/gpio/gpio-uclass.c @@ -5,6 +5,9 @@

#include <common.h> #include <dm.h> +#include <dm/device-internal.h> +#include <dm/lists.h> +#include <dm/uclass-internal.h> #include <dt-bindings/gpio/gpio.h> #include <errno.h> #include <fdtdec.h> @@ -141,6 +144,118 @@ static int gpio_find_and_xlate(struct gpio_desc

*desc,

...

return gpio_xlate_offs_flags(desc->dev, desc, args);  }

+#if defined(CONFIG_DM_GPIO_HOG)

+struct gpio_hog_priv {

• struct gpio_desc gpiod;

+};

+struct gpio_hog_data {

• int gpiod_flags;
• int value;
• u32 val[2];

+};

+static int gpio_hog_ofdata_to_platdata(struct udevice *dev) {

• struct gpio_hog_data *plat = dev_get_platdata(dev);
• const char *nodename;
• int ret;
• plat->value = 0;
• if (dev_read_bool(dev, "input")) {

• plat->gpiod_flags = GPIOD_IS_IN;
  

• } else if (dev_read_bool(dev, "output-high")) {

• plat->value = 1;
  

• plat->gpiod_flags = GPIOD_IS_OUT;
  

• } else if (dev_read_bool(dev, "output-low")) {

• plat->gpiod_flags = GPIOD_IS_OUT;
  

• } else {

• printf("%s: missing gpio-hog state.\n", __func__);
  

• return -EINVAL;
  

• }
• ret = dev_read_u32_array(dev, "gpios", plat->val, 2);
• if (ret) {

• printf("%s: wrong gpios property, 2 values needed %d\n",
  

•        __func__, ret);
  

• return ret;
  

• }
• nodename = dev_read_string(dev, "line-name");
• if (!nodename)

• nodename = dev_read_name(dev);
  

• device_set_name(dev, nodename);
• return 0;

+}

+static int gpio_hog_probe(struct udevice *dev) {

• struct gpio_hog_data *plat = dev_get_platdata(dev);
• struct gpio_hog_priv *priv = dev_get_priv(dev);
• int ret;
• ret = gpio_dev_request_index(dev->parent, dev->name, "gpio-hog",

• 		     plat->val[0], plat->gpiod_flags,
  

• 		     plat->val[1], &priv->gpiod);
  

• if (ret < 0) {

• debug("%s: node %s could not get gpio.\n", __func__,
  

•       dev->name);
  

• return ret;
  

• }
• dm_gpio_set_dir(&priv->gpiod);
• if (plat->gpiod_flags == GPIOD_IS_OUT)

• dm_gpio_set_value(&priv->gpiod, plat->value);
  

• return 0;

+}

+int gpio_hog_probe_all(void) +{

• struct udevice *dev;
• int ret;
• for (uclass_first_device(UCLASS_NOP, &dev);

•     dev;
  

•     uclass_find_next_device(&dev)) {
  

• if (dev->driver == DM_GET_DRIVER(gpio_hog)) {
  

• 	ret = device_probe(dev);
  

• 	if (ret)
  

• 		return ret;
  

• }
  

• }
• return 0;

+}

+struct gpio_desc *gpio_hog_lookup_name(const char *name) {

• struct udevice *dev;
• gpio_hog_probe_all();
• if (!uclass_get_device_by_name(UCLASS_NOP, name, &dev)) {

• struct gpio_hog_priv *priv = dev_get_priv(dev);
  

• return &priv->gpiod;
  

• }
• return NULL;

+}

+U_BOOT_DRIVER(gpio_hog) = {

• .name = "gpio_hog",
• .id = UCLASS_NOP,
• .ofdata_to_platdata = gpio_hog_ofdata_to_platdata,
• .probe = gpio_hog_probe,
• .priv_auto_alloc_size = sizeof(struct gpio_hog_priv),
• .platdata_auto_alloc_size = sizeof(struct gpio_hog_data), }; #else

+struct gpio_desc *gpio_hog_lookup_name(const char *name) {

• return NULL;

+} +#endif

int dm_gpio_request(struct gpio_desc *desc, const char *label) { struct udevice *dev = desc->dev; @@ -640,22 +755,25 @@ int dm_gpio_get_values_as_int(const struct

gpio_desc *desc_list, int count)

...

return vector; }

-static int gpio_request_tail(int ret, ofnode node, +static int gpio_request_tail(int ret, const char *nodename, struct ofnode_phandle_args *args, const char *list_name, int index,

• 	     struct gpio_desc *desc, int flags, bool add_index)
  

• 	     struct gpio_desc *desc, int flags,
  

• 	     bool add_index, struct udevice *dev)
  

{

• desc->dev = NULL;

• desc->dev = dev; desc->offset = 0; desc->flags = 0; if (ret) goto err;

• ret = uclass_get_device_by_ofnode(UCLASS_GPIO, args->node,

• 			  &desc->dev);
  

• if (ret) {

• debug("%s: uclass_get_device_by_ofnode failed\n", __func__);
  

• goto err;
  

• if (!desc->dev) {

• ret = uclass_get_device_by_ofnode(UCLASS_GPIO, args->node,
  

• 				  &desc->dev);
  

• if (ret) {
  

• 	debug("%s: uclass_get_device_by_ofnode failed\n",
  

__func__);

...

• 	goto err;
  

• }
  

  } ret = gpio_find_and_xlate(desc, args); if (ret) {

@@ -663,8 +781,7 @@ static int gpio_request_tail(int ret, ofnode node, goto err; } ret = dm_gpio_requestf(desc, add_index ? "%s.%s%d" : "%s.%s",

• 	       ofnode_get_name(node),
  

• 	       list_name, index);
  

• 	       nodename, list_name, index);
  

  if (ret) { debug("%s: dm_gpio_requestf failed\n", __func__); goto err;

@@ -678,7 +795,7 @@ static int gpio_request_tail(int ret, ofnode node, return 0; err: debug("%s: Node '%s', property '%s', failed to request GPIO index %d:

%d\n",

...

•      __func__, ofnode_get_name(node), list_name, index, ret);
  

•      __func__, nodename, list_name, index, ret);
  

  return ret;

}

@@ -692,8 +809,8 @@ static int _gpio_request_by_name_nodev(ofnode node,

const char *list_name,

...

ret = ofnode_parse_phandle_with_args(node, list_name, "#gpio-cells", 0, index, &args);

• return gpio_request_tail(ret, node, &args, list_name, index, desc,

• 		 flags, add_index);
  

• return gpio_request_tail(ret, ofnode_get_name(node), &args, list_name,

• 		 index, desc, flags, add_index, NULL);
  

}

int gpio_request_by_name_nodev(ofnode node, const char *list_name, int index, @@ -707,13 +824,14 @@ int gpio_request_by_name(struct udevice

*dev, const char *list_name, int index,

...

	 struct gpio_desc *desc, int flags)  {

struct ofnode_phandle_args args;

• ofnode node; int ret;

  ret = dev_read_phandle_with_args(dev, list_name, "#gpio-cells", 0, index, &args);

• return gpio_request_tail(ret, dev_ofnode(dev), &args, list_name,

• 		 index, desc, flags, index > 0);
  

• node = dev_ofnode(dev);
• return gpio_request_tail(ret, ofnode_get_name(node), &args, list_name,

• 		 index, desc, flags, index > 0, NULL);
  

}

int gpio_request_list_by_name_nodev(ofnode node, const char *list_name, @@ -854,8 +972,28 @@ static int gpio_pre_remove(struct udevice

*dev)

...

return gpio_renumber(dev); }

+int gpio_dev_request_index(struct udevice *dev, const char *nodename,

• 	   char *list_name, int index, int flags,
  

• 	   int dtflags, struct gpio_desc *desc) {
  

• struct ofnode_phandle_args args;
• args.node = ofnode_null();
• args.args_count = 2;
• args.args[0] = index;
• args.args[1] = dtflags;
• return gpio_request_tail(0, nodename, &args, list_name, index, desc,

• 		 flags, 0, dev);
  

+}

static int gpio_post_bind(struct udevice *dev) { +#if defined(CONFIG_DM_GPIO_HOG)

• struct udevice *child;
• ofnode node;

+#endif

#if defined(CONFIG_NEEDS_MANUAL_RELOC) struct dm_gpio_ops *ops = (struct dm_gpio_ops *)device_get_ops(dev); static int reloc_done; @@ -885,6 +1023,17 @@ static int gpio_post_bind(struct udevice *dev) reloc_done++; } #endif

+#if defined(CONFIG_DM_GPIO_HOG)

• dev_for_each_subnode(node, dev) {

• if (ofnode_read_bool(node, "gpio-hog")) {
  

• 	const char *name = ofnode_get_name(node);
  

• 	device_bind_driver_to_node(dev, "gpio_hog", name,
  

• 				   node, &child);
  

• }
  

• }

+#endif return 0; }

diff --git a/include/asm-generic/gpio.h b/include/asm-generic/gpio.h index d03602696f..37f71e5708 100644 --- a/include/asm-generic/gpio.h +++ b/include/asm-generic/gpio.h @@ -348,6 +348,22 @@ const char *gpio_get_bank_info(struct udevice *dev,

int *offset_count);

...

*/ int dm_gpio_lookup_name(const char *name, struct gpio_desc *desc);

+/**

• • gpio_hog_lookup_name() - Look up a named GPIO and return the gpio

descr.

...

• • @name: Name to look up
• • @return: Returns gpio_desc for gpio
• */

+struct gpio_desc *gpio_hog_lookup_name(const char *name);

+/**

• • gpio_hog_probe_all() - probe all gpio devices with
• • gpio-hog subnodes.
• • @return: Returns return value from device_probe()
• */

+int gpio_hog_probe_all(void);

/**

• gpio_lookup_name - Look up a GPIO name and return its details

@@ -503,6 +519,22 @@ int gpio_request_list_by_name_nodev(ofnode node,

const char *list_name,

...

		    struct gpio_desc *desc_list, int max_count,
		    int flags);

+/**

• • gpio_dev_request_index() - request single GPIO from gpio device
• • @dev: GPIO device
• • @nodename: Name of node
• • @list_name: Name of GPIO list (e.g. "board-id-gpios")
• • @index: Index number of the GPIO in that list use request (0=first)
• • @flags: GPIOD_* flags
• • @dtflags: GPIO flags read from DT
• • @desc: GPIO descriotor filled from this function
• • @return: return value from gpio_request_tail()
• */

+int gpio_dev_request_index(struct udevice *dev, const char *nodename,

• 	   char *list_name, int index, int flags,
  

• 	   int dtflags, struct gpio_desc *desc);
  

/**

• dm_gpio_free() - Free a single GPIO

Tested-by: Michal Simek michal.simek@xilinx.com (zcu102)

Thanks, Michal

For the patch Review and tested on stm32mp1 board.

Tested-by: Patrick Delaunay patrick.delaunay@st.com

Regards PAtrick

Hi Heiko,

From: Heiko Schocher hs@denx.de Sent: samedi 22 juin 2019 07:32

Hello Patrick,

Am 21.06.2019 um 18:09 schrieb Patrick DELAUNAY:

...

Hi Mickael

...

Subject: Re: [PATCH v4] gpio: add gpio-hog support Importance: High

On 12. 06. 19 6:11, Heiko Schocher wrote:

...

add gpio-hog support. GPIO hogging is a mechanism providing automatic GPIO request and configuration as part of the gpio-controller's driver probe function.

for more infos see: doc/device-tree-bindings/gpio/gpio.txt

Signed-off-by: Heiko Schocher hs@denx.de

---

clean travis build, see result: https://travis-ci.org/hsdenx/u-boot-test/builds/544040419

Changes in v4:

• rework as suggested by Patrick based on patch: http://patchwork.ozlabs.org/patch/1098913/ use new UCLASS_NOP hog detection + DT parsing in gpio_post_bind() .... info saved in platdata gpio configuration for hog (request and set_dir) => probe

function

Changes in v3:

• probe now all gpio devices with gpio-hogs from board_r before board_late_init() call or if someone wants to access gpio-hog based gpios with gpio_hog_lookup_name() This is needed, because we cannot be sure, if a gpio device which contains gpio-hogs is probed.
• add line-name property as Michal recommended
• renamed gpio_priv_one into gpio_priv_one_hog and protected through CONFIG_DM_GPIO_HOG

Changes in v2:

• move gpio_hog() call from post_probe() to post_bind() call. Check if current gpio device has gpio-hog definitions, if so, probe it.

  common/board_r.c | 6 + doc/device-tree-bindings/gpio/gpio.txt | 55 ++++++++ drivers/gpio/Kconfig | 10 ++ drivers/gpio/gpio-uclass.c | 181 ++++++++++++++++++++++--- include/asm-generic/gpio.h | 32 +++++ 5 files changed, 268 insertions(+), 16 deletions(-)

diff --git a/common/board_r.c b/common/board_r.c index 150e8cd424..507da4c74a 100644 --- a/common/board_r.c +++ b/common/board_r.c @@ -49,6 +49,9 @@ #include <linux/err.h> #include <efi_loader.h> #include <wdt.h> +#if defined(CONFIG_DM_GPIO_HOG) +#include <asm/gpio.h> +#endif

DECLARE_GLOBAL_DATA_PTR;

@@ -796,6 +799,9 @@ static init_fnc_t init_sequence_r[] = { #ifdef CONFIG_CMD_NET initr_ethaddr, #endif +#if defined(CONFIG_DM_GPIO_HOG)

• gpio_hog_probe_all,

+#endif

Any reason to have hog at this place ?

Not really, but we need to check at some point if all gpio hogs are processed, and I thought at a late time it does not break anything...

Ok. It is safe....

and anyway any board can choose the call the function earlier if the hog was required (for nay initcall before board_late_init : initr_nand, initr_mmc ....)

...

Before board_late_init....

I expect more before board_init() for my future use-case, to replace the current (dirty) implementation with pinctrl-0

#if defined(CONFIG_WDT) initr_watchdog, #endif +#if defined(CONFIG_DM_GPIO_HOG)

• gpio_hog_probe_all,

+#endif #if defined(CONFIG_ARM) || defined(CONFIG_NDS32) ||

defined(CONFIG_RISCV) || \

...

defined(CONFIG_SANDBOX) board_init, /* Setup chipselects */ #endif

In my device tree I have :

stmfx: stmfx@42 { compatible = "st,stmfx-0300"; reg = <0x42>; interrupts = <8 IRQ_TYPE_EDGE_RISING>; interrupt-parent = <&gpioi>; vdd-supply = <&v3v3>;

stmfx_pinctrl: stmfx-pin-controller {
	compatible = "st,stmfx-0300-pinctrl";
	gpio-controller;
	#gpio-cells = <2>;
	interrupt-controller;
	#interrupt-cells = <2>;
	gpio-ranges = <&stmfx_pinctrl 0 0 24>;
	pinctrl-names = "default";
	pinctrl-0 = <&hog_pins>;

	hog_pins: hog {
		pins = "gpio14";
		drive-push-pull;
		bias-pull-down;
	};

};

And in board, I force probe to configure HOG....before to initialize the DISPLAY, so in

./board/st/stm32mp1/stm32mp1.c : board _init()

/* board dependent setup after realloc */ int board_init(void) { ..... /* probe all PINCTRL for hog */ for (uclass_first_device(UCLASS_PINCTRL, &dev); dev; uclass_next_device(&dev)) { pr_debug("probe pincontrol = %s\n", dev->name); }

Or you can let each board choose when call the function as it is already done for other function as regulators_enable_boot_on() So don't add it in init_sequence_r[] / board_r.c ?...

Hmm... I do not like this, because if you enable gpio-hog support, you don;t want to think to call gpio_hog_probe_all().

But it should be possible to call gpio_hog_probe_all() earlier from board code, as gpio_hog_probe_all() calls device_probe():

Fine.

212 int gpio_hog_probe_all(void) 213 { 214 struct udevice *dev; 215 int ret; 216 217 for (uclass_first_device(UCLASS_NOP, &dev); 218 dev; 219 uclass_find_next_device(&dev)) { 220 if (dev->driver == DM_GET_DRIVER(gpio_hog)) { 221 ret = device_probe(dev);

device_probe() checks if it is already probed, and simply returns if so, see:

http://git.denx.de/?p=u- boot.git;a=blob;f=drivers/core/device.c;h=0d15e5062b66123cd364bd9803e076db7 e7dd97c;hb=77f6e2dd0551d8a825bab391a1bd6b838874bcd4#l319

May you can test if this works for you?

I have the same idea, it is why I say is was no't blocking for me.

Already tested and it is working (gpio hog is called only one time even if gpio_hog_probe_all several time).

...

Anyway it is not blocking for me, it just a question.

Fine.

...

...

...

#ifdef CONFIG_BOARD_LATE_INIT board_late_init,

[...]

...

For the patch Review and tested on stm32mp1 board.

Tested-by: Patrick Delaunay patrick.delaunay@st.com

Thanks for testing!

bye, Heiko

Regards Patrick

Hello Simon,

Am 22.06.2019 um 21:09 schrieb Simon Glass:

Hi Heiko,

On Wed, 12 Jun 2019 at 05:12, Heiko Schocher hs@denx.de wrote:

...

add gpio-hog support. GPIO hogging is a mechanism providing automatic GPIO request and configuration as part of the gpio-controller's driver probe function.

for more infos see: doc/device-tree-bindings/gpio/gpio.txt

Signed-off-by: Heiko Schocher hs@denx.de

---

clean travis build, see result: https://travis-ci.org/hsdenx/u-boot-test/builds/544040419

Changes in v4:

• rework as suggested by Patrick based on patch: http://patchwork.ozlabs.org/patch/1098913/ use new UCLASS_NOP hog detection + DT parsing in gpio_post_bind() .... info saved in platdata gpio configuration for hog (request and set_dir) => probe function

Sorry for not getting to this earlier.

no problem, thanks for your time!

...

Changes in v3:

• probe now all gpio devices with gpio-hogs from board_r before board_late_init() call or if someone wants to access gpio-hog based gpios with gpio_hog_lookup_name() This is needed, because we cannot be sure, if a gpio device which contains gpio-hogs is probed.
• add line-name property as Michal recommended
• renamed gpio_priv_one into gpio_priv_one_hog and protected through CONFIG_DM_GPIO_HOG

Changes in v2:

• move gpio_hog() call from post_probe() to post_bind() call. Check if current gpio device has gpio-hog definitions, if so, probe it.

  common/board_r.c | 6 + doc/device-tree-bindings/gpio/gpio.txt | 55 ++++++++ drivers/gpio/Kconfig | 10 ++ drivers/gpio/gpio-uclass.c | 181 ++++++++++++++++++++++--- include/asm-generic/gpio.h | 32 +++++ 5 files changed, 268 insertions(+), 16 deletions(-)

diff --git a/common/board_r.c b/common/board_r.c index 150e8cd424..507da4c74a 100644 --- a/common/board_r.c +++ b/common/board_r.c @@ -49,6 +49,9 @@ #include <linux/err.h> #include <efi_loader.h> #include <wdt.h> +#if defined(CONFIG_DM_GPIO_HOG) +#include <asm/gpio.h> +#endif

DECLARE_GLOBAL_DATA_PTR;

@@ -796,6 +799,9 @@ static init_fnc_t init_sequence_r[] = { #ifdef CONFIG_CMD_NET initr_ethaddr, #endif +#if defined(CONFIG_DM_GPIO_HOG)

•   gpio_hog_probe_all,
  

+#endif #ifdef CONFIG_BOARD_LATE_INIT board_late_init, #endif diff --git a/doc/device-tree-bindings/gpio/gpio.txt b/doc/device-tree-bindings/gpio/gpio.txt index f7a158d858..e774439369 100644 --- a/doc/device-tree-bindings/gpio/gpio.txt +++ b/doc/device-tree-bindings/gpio/gpio.txt @@ -210,3 +210,58 @@ Example 2: Here, three GPIO ranges are defined wrt. two pin controllers. pinctrl1 GPIO ranges are defined using pin numbers whereas the GPIO ranges wrt. pinctrl2 are named "foo" and "bar".

+3) GPIO hog definitions +-----------------------

+The GPIO chip may contain GPIO hog definitions. GPIO hogging is a mechanism +providing automatic GPIO request and configuration as part of the +gpio-controller's driver probe function.

+Each GPIO hog definition is represented as a child node of the GPIO controller. +Required properties: +- gpio-hog: A property specifying that this child node represents a GPIO hog. +- gpios: Store the GPIO information (id, flags) for the GPIO to

•         affect.
  

•          ! Not yet support more than one gpio !
  

+Only one of the following properties scanned in the order shown below. +- input: A property specifying to set the GPIO direction as input. +- output-low A property specifying to set the GPIO direction as output with

•         the value low.
  

+- output-high A property specifying to set the GPIO direction as output with

•         the value high.
  

+Optional properties: +- line-name: The GPIO label name. If not present the node name is used.

This seems a strange name. I suppose this is what linux called it.

Yes.

...

+Example:

•    tca6416@20 {
  

•            compatible = "ti,tca6416";
  

•            reg = <0x20>;
  

•            #gpio-cells = <2>;
  

•            gpio-controller;
  

•            env_reset {
  

•                    gpio-hog;
  

•                    input;
  

•                    gpios = <6 GPIO_ACTIVE_LOW>;
  

•            };
  

•            boot_rescue {
  

•                    gpio-hog;
  

•                    input;
  

•                    gpios = <7 GPIO_ACTIVE_LOW>;
  

Should have an example for line-name

Oh, yes of course, I added the "line-name" part later, and forgot the example.

...

•            };
  

•    };
  

+For the above Example you can than access the gpio in your boardcode +with:

•    desc = gpio_hog_lookup_name("boot_rescue.gpio-hog");
  

Could we drop the .gpio-hog suffix? It doesn't seem necessary, but perhaps you are worried about people getting confused otherwise?

Good catch! Since v4 it is already without this suffix, so I correct the doc here.

Also this function should be passed a desc struct pointer, not return static data.

...

•    if (desc) {
  

•            if (dm_gpio_get_value(desc))
  

You should check the error here.

dm_gpio_get_value() returns -ve on error, including if desc returns an invalid GPIO (dev=NULL), so you should be able to do:

gpio_hog_lookup_name("boot_rescue.gpio-hog", &desc);

Also we should check here the return value.

if (dm_gpio_get_value(desc) == 1)

Good point. I rework this, thanks!

So in fact

...

•                    printf("\nBooting into Rescue System\n");
  

•           else
  

•                   printf("\nBoot normal\n");
  

diff --git a/drivers/gpio/Kconfig b/drivers/gpio/Kconfig index e36a8abc42..fa1c99700f 100644 --- a/drivers/gpio/Kconfig +++ b/drivers/gpio/Kconfig @@ -14,6 +14,16 @@ config DM_GPIO particular GPIOs that they provide. The uclass interface is defined in include/asm-generic/gpio.h.

+config DM_GPIO_HOG

Can we drop the DM_ prefix? At some point DM_GPIO will go away and we don't want to rename these sorts of options.

Hmm... we also have the DM_GPIO symbol, as the gpio hog only works with DM_GPIO enabled, I thought DM_GPIO_HOG is the better choice, but of course I can change this.

...

•   bool "Enable GPIO hog support"
  

•   depends on DM_GPIO
  

•   default n
  

•   help
  

•     Enable gpio hog support
  

•     The GPIO chip may contain GPIO hog definitions. GPIO hogging
  

•     is a mechanism providing automatic GPIO request and config-
  

•     uration as part of the gpio-controller's driver probe function.
  

• config ALTERA_PIO bool "Altera PIO driver" depends on DM_GPIO

diff --git a/drivers/gpio/gpio-uclass.c b/drivers/gpio/gpio-uclass.c index da5e9ba6e5..308d0863ad 100644 --- a/drivers/gpio/gpio-uclass.c +++ b/drivers/gpio/gpio-uclass.c @@ -5,6 +5,9 @@

#include <common.h> #include <dm.h> +#include <dm/device-internal.h> +#include <dm/lists.h> +#include <dm/uclass-internal.h> #include <dt-bindings/gpio/gpio.h> #include <errno.h> #include <fdtdec.h> @@ -141,6 +144,118 @@ static int gpio_find_and_xlate(struct gpio_desc *desc, return gpio_xlate_offs_flags(desc->dev, desc, args); }

+#if defined(CONFIG_DM_GPIO_HOG)

+struct gpio_hog_priv {

•   struct gpio_desc gpiod;
  

As above I don't think you need this.

Hmm.. when probing I need to request the gpio with gpio_dev_request_index() for which I need to pass a pointer to struct gpio_desc ...

And later I want to have the possibility to access the gpio for example with dm_gpio_set_value() and need therefore the gpio descriptor, so I must store it somewhere. Or do I overlook something obvious?

...

+};

+struct gpio_hog_data {

•   int gpiod_flags;
  

•   int value;
  

•   u32 val[2];
  

+};

+static int gpio_hog_ofdata_to_platdata(struct udevice *dev) +{

•   struct gpio_hog_data *plat = dev_get_platdata(dev);
  

•   const char *nodename;
  

•   int ret;
  

•   plat->value = 0;
  

•   if (dev_read_bool(dev, "input")) {
  

•           plat->gpiod_flags = GPIOD_IS_IN;
  

•   } else if (dev_read_bool(dev, "output-high")) {
  

•           plat->value = 1;
  

•           plat->gpiod_flags = GPIOD_IS_OUT;
  

•   } else if (dev_read_bool(dev, "output-low")) {
  

•           plat->gpiod_flags = GPIOD_IS_OUT;
  

•   } else {
  

•           printf("%s: missing gpio-hog state.\n", __func__);
  

•           return -EINVAL;
  

•   }
  

•   ret = dev_read_u32_array(dev, "gpios", plat->val, 2);
  

•   if (ret) {
  

•           printf("%s: wrong gpios property, 2 values needed %d\n",
  

•                  __func__, ret);
  

•           return ret;
  

•   }
  

•   nodename = dev_read_string(dev, "line-name");
  

•   if (!nodename)
  

•           nodename = dev_read_name(dev);
  

•   device_set_name(dev, nodename);
  

I don't think you should set the device name unless needed. But why not store this in the device data, rather than changing its name? If we want to change the name, why not just use that for the node name in the .dts?

Because I use the device name for finding the gpio through gpio_hog_lookup_name().

Ok, if there is no line-name property, we do not need to set the device name, as it has already the correct name, so this can be changed to:

nodename = dev_read_string(dev, "line-name"); if (nodename) device_set_name(dev, nodename);

...

•   return 0;
  

+}

+static int gpio_hog_probe(struct udevice *dev) +{

•   struct gpio_hog_data *plat = dev_get_platdata(dev);
  

•   struct gpio_hog_priv *priv = dev_get_priv(dev);
  

•   int ret;
  

•   ret = gpio_dev_request_index(dev->parent, dev->name, "gpio-hog",
  

•                                plat->val[0], plat->gpiod_flags,
  

•                                plat->val[1], &priv->gpiod);
  

•   if (ret < 0) {
  

•           debug("%s: node %s could not get gpio.\n", __func__,
  

•                 dev->name);
  

•           return ret;
  

•   }
  

•   dm_gpio_set_dir(&priv->gpiod);
  

Check error

Yes ... Hmm, while checking, this is not needed anymore, as gpio_dev_request_index() calls gpio_request_tail() which calls dm_gpio_set_dir(), so I remove this.

...

•   if (plat->gpiod_flags == GPIOD_IS_OUT)
  

•           dm_gpio_set_value(&priv->gpiod, plat->value);
  

check error

done.

...

•   return 0;
  

+}

+int gpio_hog_probe_all(void) +{

•   struct udevice *dev;
  

•   int ret;
  

•   for (uclass_first_device(UCLASS_NOP, &dev);
  

Perhaps you should create a new uclass for this?

Hmm.. yes, I had an approach with a new uclass, but I thought, the new [1] UCLASS_NOP is valid for this here to?

[1] http://patchwork.ozlabs.org/patch/1098913/

...

•        dev;
  

•        uclass_find_next_device(&dev)) {
  

•           if (dev->driver == DM_GET_DRIVER(gpio_hog)) {
  

•                   ret = device_probe(dev);
  

•                   if (ret)
  

•                           return ret;
  

•           }
  

•   }
  

•   return 0;
  

+}

+struct gpio_desc *gpio_hog_lookup_name(const char *name) +{

•   struct udevice *dev;
  

•   gpio_hog_probe_all();
  

•   if (!uclass_get_device_by_name(UCLASS_NOP, name, &dev)) {
  

•           struct gpio_hog_priv *priv = dev_get_priv(dev);
  

•           return &priv->gpiod;
  

•   }
  

•   return NULL;
  

+}

+U_BOOT_DRIVER(gpio_hog) = {

•   .name   = "gpio_hog",
  

•   .id     = UCLASS_NOP,
  

•   .ofdata_to_platdata = gpio_hog_ofdata_to_platdata,
  

•   .probe = gpio_hog_probe,
  

•   .priv_auto_alloc_size = sizeof(struct gpio_hog_priv),
  

•   .platdata_auto_alloc_size = sizeof(struct gpio_hog_data),
  

+}; +#else +struct gpio_desc *gpio_hog_lookup_name(const char *name) +{

•   return NULL;
  

+} +#endif

• int dm_gpio_request(struct gpio_desc *desc, const char *label) { struct udevice *dev = desc->dev;

@@ -640,22 +755,25 @@ int dm_gpio_get_values_as_int(const struct gpio_desc *desc_list, int count) return vector; }

-static int gpio_request_tail(int ret, ofnode node, +static int gpio_request_tail(int ret, const char *nodename, struct ofnode_phandle_args *args, const char *list_name, int index,

•                        struct gpio_desc *desc, int flags, bool add_index)
  

•                        struct gpio_desc *desc, int flags,
  

•                        bool add_index, struct udevice *dev)
  

This function badly needs a comment now. Also 'dev' is not a great name.

Before my patch I think, it needed a comment ;-) I try to add one...

Hmm... the name "dev" is the same as in struct gpio_desc ... what is a good name?

parent? gpio_dev?

I tend to name it gpio_dev ...

...

{

•   desc->dev = NULL;
  

•   desc->dev = dev;
     desc->offset = 0;
     desc->flags = 0;
     if (ret)
             goto err;
  
  

•   ret = uclass_get_device_by_ofnode(UCLASS_GPIO, args->node,
  

•                                     &desc->dev);
  

•   if (ret) {
  

•           debug("%s: uclass_get_device_by_ofnode failed\n", __func__);
  

•           goto err;
  

•   if (!desc->dev) {
  

•           ret = uclass_get_device_by_ofnode(UCLASS_GPIO, args->node,
  

•                                             &desc->dev);
  

•           if (ret) {
  

•                   debug("%s: uclass_get_device_by_ofnode failed\n", __func__);
  

•                   goto err;
  

•           }
     }
     ret = gpio_find_and_xlate(desc, args);
     if (ret) {
  

@@ -663,8 +781,7 @@ static int gpio_request_tail(int ret, ofnode node, goto err; } ret = dm_gpio_requestf(desc, add_index ? "%s.%s%d" : "%s.%s",

•                          ofnode_get_name(node),
  

•                          list_name, index);
  

•                          nodename, list_name, index);
     if (ret) {
             debug("%s: dm_gpio_requestf failed\n", __func__);
             goto err;
  

@@ -678,7 +795,7 @@ static int gpio_request_tail(int ret, ofnode node, return 0; err: debug("%s: Node '%s', property '%s', failed to request GPIO index %d: %d\n",

•         __func__, ofnode_get_name(node), list_name, index, ret);
  

•         __func__, nodename, list_name, index, ret);
     return ret;
  

  }

@@ -692,8 +809,8 @@ static int _gpio_request_by_name_nodev(ofnode node, const char *list_name, ret = ofnode_parse_phandle_with_args(node, list_name, "#gpio-cells", 0, index, &args);

•   return gpio_request_tail(ret, node, &args, list_name, index, desc,
  

•                            flags, add_index);
  

•   return gpio_request_tail(ret, ofnode_get_name(node), &args, list_name,
  

•                            index, desc, flags, add_index, NULL);
  

  }

  int gpio_request_by_name_nodev(ofnode node, const char *list_name, int index,

@@ -707,13 +824,14 @@ int gpio_request_by_name(struct udevice *dev, const char *list_name, int index, struct gpio_desc *desc, int flags) { struct ofnode_phandle_args args;

•   ofnode node;
     int ret;
  
     ret = dev_read_phandle_with_args(dev, list_name, "#gpio-cells", 0,
                                      index, &args);
  

•   return gpio_request_tail(ret, dev_ofnode(dev), &args, list_name,
  

•                            index, desc, flags, index > 0);
  

•   node = dev_ofnode(dev);
  

•   return gpio_request_tail(ret, ofnode_get_name(node), &args, list_name,
  

•                            index, desc, flags, index > 0, NULL);
  

  }

  int gpio_request_list_by_name_nodev(ofnode node, const char *list_name,

@@ -854,8 +972,28 @@ static int gpio_pre_remove(struct udevice *dev) return gpio_renumber(dev); }

+int gpio_dev_request_index(struct udevice *dev, const char *nodename,

•                      char *list_name, int index, int flags,
  

•                      int dtflags, struct gpio_desc *desc)
  

+{

•   struct ofnode_phandle_args args;
  

•   args.node =  ofnode_null();
  

•   args.args_count = 2;
  

•   args.args[0] = index;
  

•   args.args[1] = dtflags;
  

•   return gpio_request_tail(0, nodename, &args, list_name, index, desc,
  

•                            flags, 0, dev);
  

+}

• static int gpio_post_bind(struct udevice *dev) {

+#if defined(CONFIG_DM_GPIO_HOG)

•   struct udevice *child;
  

•   ofnode node;
  

+#endif

• #if defined(CONFIG_NEEDS_MANUAL_RELOC) struct dm_gpio_ops *ops = (struct dm_gpio_ops *)device_get_ops(dev); static int reloc_done;

@@ -885,6 +1023,17 @@ static int gpio_post_bind(struct udevice *dev) reloc_done++; } #endif

+#if defined(CONFIG_DM_GPIO_HOG)

Can you use if (IS_ENABLED(CONFIG....

Yes, changed.

...

•   dev_for_each_subnode(node, dev) {
  

•           if (ofnode_read_bool(node, "gpio-hog")) {
  

•                   const char *name = ofnode_get_name(node);
  

•                   device_bind_driver_to_node(dev, "gpio_hog", name,
  

•                                              node, &child);
  

Check error? This might run out of memory, perhaps.

done.

...

•           }
  

•   }
  

+#endif return 0; }

diff --git a/include/asm-generic/gpio.h b/include/asm-generic/gpio.h index d03602696f..37f71e5708 100644 --- a/include/asm-generic/gpio.h +++ b/include/asm-generic/gpio.h @@ -348,6 +348,22 @@ const char *gpio_get_bank_info(struct udevice *dev, int *offset_count); */ int dm_gpio_lookup_name(const char *name, struct gpio_desc *desc);

+/**

• • gpio_hog_lookup_name() - Look up a named GPIO and return the gpio descr.
• • @name: Name to look up
• • @return: Returns gpio_desc for gpio
• */

+struct gpio_desc *gpio_hog_lookup_name(const char *name);

+/**

• • gpio_hog_probe_all() - probe all gpio devices with
• • gpio-hog subnodes.
• • @return: Returns return value from device_probe()

What if one fails and the others succeed?

Currently I return with error code from device_probe(). So the rest of the gpio-hogs are not probed.

Hmm.. may it is better to emmit a debug message and continue?

...

• */

+int gpio_hog_probe_all(void);

• /**
  • gpio_lookup_name - Look up a GPIO name and return its details

@@ -503,6 +519,22 @@ int gpio_request_list_by_name_nodev(ofnode node, const char *list_name, struct gpio_desc *desc_list, int max_count, int flags);

+/**

• • gpio_dev_request_index() - request single GPIO from gpio device
• • @dev: GPIO device
• • @nodename: Name of node

What is a node? Can you expand this a bit?

changed to:

* @nodename: Name of node for which gpio gets requested, used * for the gpio label name

...

• • @list_name: Name of GPIO list (e.g. "board-id-gpios")
• • @index: Index number of the GPIO in that list use request (0=first)
• • @flags: GPIOD_* flags
• • @dtflags: GPIO flags read from DT

Reference GPIOD_... mask here?

Yep, added.

...

• • @desc: GPIO descriotor filled from this function

returns GPIO descriptor...

changed.

...

• • @return: return value from gpio_request_tail()
• */

+int gpio_dev_request_index(struct udevice *dev, const char *nodename,

•                      char *list_name, int index, int flags,
  

•                      int dtflags, struct gpio_desc *desc);
  

• /**
  • dm_gpio_free() - Free a single GPIO

-- 2.21.0

Regards, Simon

Thanks for this review!

bye, Heiko

Hello Simon,

Am 26.06.2019 um 17:07 schrieb Simon Glass:

Hi Heiko,

On Mon, 24 Jun 2019 at 00:16, Heiko Schocher hs@denx.de wrote:

...

Hello Simon,

Am 22.06.2019 um 21:09 schrieb Simon Glass:

...

Hi Heiko,

On Wed, 12 Jun 2019 at 05:12, Heiko Schocher hs@denx.de wrote:

...

add gpio-hog support. GPIO hogging is a mechanism providing automatic GPIO request and configuration as part of the gpio-controller's driver probe function.

for more infos see: doc/device-tree-bindings/gpio/gpio.txt

Signed-off-by: Heiko Schocher hs@denx.de

---

clean travis build, see result: https://travis-ci.org/hsdenx/u-boot-test/builds/544040419

Changes in v4:

• rework as suggested by Patrick based on patch: http://patchwork.ozlabs.org/patch/1098913/ use new UCLASS_NOP hog detection + DT parsing in gpio_post_bind() .... info saved in platdata gpio configuration for hog (request and set_dir) => probe function

[..]

...

...

...

+config DM_GPIO_HOG

Can we drop the DM_ prefix? At some point DM_GPIO will go away and we don't want to rename these sorts of options.

Hmm... we also have the DM_GPIO symbol, as the gpio hog only works with DM_GPIO enabled, I thought DM_GPIO_HOG is the better choice, but of course I can change this.

Yes I understand that, but in general we hope one day to rename DM_GPIO to GPIO, once everything uses DM. That's why I am trying to keep the DM_ suffix only for things that need it.

Yep, already dropped, see temp version (travisbuild clean): https://github.com/hsdenx/u-boot-test/commit/509bad758b8abc94dbbebc2a387b9e9...

...

...

...

•   bool "Enable GPIO hog support"
  

•   depends on DM_GPIO
  

•   default n
  

•   help
  

•     Enable gpio hog support
  

•     The GPIO chip may contain GPIO hog definitions. GPIO hogging
  

•     is a mechanism providing automatic GPIO request and config-
  

•     uration as part of the gpio-controller's driver probe function.
  

• config ALTERA_PIO bool "Altera PIO driver" depends on DM_GPIO

diff --git a/drivers/gpio/gpio-uclass.c b/drivers/gpio/gpio-uclass.c index da5e9ba6e5..308d0863ad 100644 --- a/drivers/gpio/gpio-uclass.c +++ b/drivers/gpio/gpio-uclass.c @@ -5,6 +5,9 @@

#include <common.h> #include <dm.h> +#include <dm/device-internal.h> +#include <dm/lists.h> +#include <dm/uclass-internal.h> #include <dt-bindings/gpio/gpio.h> #include <errno.h> #include <fdtdec.h> @@ -141,6 +144,118 @@ static int gpio_find_and_xlate(struct gpio_desc *desc, return gpio_xlate_offs_flags(desc->dev, desc, args); }

+#if defined(CONFIG_DM_GPIO_HOG)

+struct gpio_hog_priv {

•   struct gpio_desc gpiod;
  

As above I don't think you need this.

Hmm.. when probing I need to request the gpio with gpio_dev_request_index() for which I need to pass a pointer to struct gpio_desc ...

The caller can declare this in a local variable.

The caller is always the device probe function, as gpio_hog_probe_all() loops over all gpio hog devices and probes them with device_probe() which cannot return a gpio_desc pointer, nor can such a pointer passed to.

Why isn't it valid to store in each gpio hog device the gpio desc it belongs too in device private data?

...

And later I want to have the possibility to access the gpio for example with dm_gpio_set_value() and need therefore the gpio descriptor, so I must store it somewhere. Or do I overlook something obvious?

In that case the caller should store it, perhaps in a local variable, or perhaps in a private data structure if needed for a long time.

[..]

Of course I can introduce a list for each gpio hog definition filled when calling gpio_hog_probe_all() ... but than I have a second list which uses memory, may has to keep in sync the list if device are removed ... I want to avoid such an approach ... as we have already a list of gpio hog devices and the priv storage pointer for device private data.

What do you think?

...

...

...

•   return 0;
  

+}

+int gpio_hog_probe_all(void) +{

•   struct udevice *dev;
  

•   int ret;
  

•   for (uclass_first_device(UCLASS_NOP, &dev);
  

Perhaps you should create a new uclass for this?

Hmm.. yes, I had an approach with a new uclass, but I thought, the new [1] UCLASS_NOP is valid for this here to?

[1] http://patchwork.ozlabs.org/patch/1098913/

I see your point and I can see arguments for each side. But the fact that you have to check the driver suggests to me that a new uclass is better.

[..]

Ok with this approach the gpio_hog additions to gpio-uclass can be moved to a gpio-hog-uclass.c file. Also the problem where to store the gpio descriptor moves from device private data to uclass private data?

Is it worth?

...

...

...

+/**

• • gpio_hog_probe_all() - probe all gpio devices with
• • gpio-hog subnodes.
• • @return: Returns return value from device_probe()

What if one fails and the others succeed?

Currently I return with error code from device_probe(). So the rest of the gpio-hogs are not probed.

Hmm.. may it is better to emmit a debug message and continue?

Sounds reasonable. But it seems to me that if there is an error there should be some visible report, so this could perhaps return the error code? Then the caller could print a messag?

Ok, so I add a printf message. I like more to add it in gpio_hog_probe_all(), because more than one gpio_hog can fail.

bye, Heiko

Hi Heiko,

On Wed, 26 Jun 2019 at 22:12, Heiko Schocher hs@denx.de wrote:

Hello Simon,

Am 26.06.2019 um 17:07 schrieb Simon Glass:

...

Hi Heiko,

On Mon, 24 Jun 2019 at 00:16, Heiko Schocher hs@denx.de wrote:

...

Hello Simon,

Am 22.06.2019 um 21:09 schrieb Simon Glass:

...

Hi Heiko,

On Wed, 12 Jun 2019 at 05:12, Heiko Schocher hs@denx.de wrote:

...

add gpio-hog support. GPIO hogging is a mechanism providing automatic GPIO request and configuration as part of the gpio-controller's driver probe function.

for more infos see: doc/device-tree-bindings/gpio/gpio.txt

Signed-off-by: Heiko Schocher hs@denx.de

---

clean travis build, see result: https://travis-ci.org/hsdenx/u-boot-test/builds/544040419

Changes in v4:

• rework as suggested by Patrick based on patch: http://patchwork.ozlabs.org/patch/1098913/ use new UCLASS_NOP hog detection + DT parsing in gpio_post_bind() .... info saved in platdata gpio configuration for hog (request and set_dir) => probe function

[..]

...

...

...

+config DM_GPIO_HOG

Can we drop the DM_ prefix? At some point DM_GPIO will go away and we don't want to rename these sorts of options.

Hmm... we also have the DM_GPIO symbol, as the gpio hog only works with DM_GPIO enabled, I thought DM_GPIO_HOG is the better choice, but of course I can change this.

Yes I understand that, but in general we hope one day to rename DM_GPIO to GPIO, once everything uses DM. That's why I am trying to keep the DM_ suffix only for things that need it.

Yep, already dropped, see temp version (travisbuild clean): https://github.com/hsdenx/u-boot-test/commit/509bad758b8abc94dbbebc2a387b9e9...

...

...

...

...

•   bool "Enable GPIO hog support"
  

•   depends on DM_GPIO
  

•   default n
  

•   help
  

•     Enable gpio hog support
  

•     The GPIO chip may contain GPIO hog definitions. GPIO hogging
  

•     is a mechanism providing automatic GPIO request and config-
  

•     uration as part of the gpio-controller's driver probe function.
  

• config ALTERA_PIO bool "Altera PIO driver" depends on DM_GPIO

diff --git a/drivers/gpio/gpio-uclass.c b/drivers/gpio/gpio-uclass.c index da5e9ba6e5..308d0863ad 100644 --- a/drivers/gpio/gpio-uclass.c +++ b/drivers/gpio/gpio-uclass.c @@ -5,6 +5,9 @@

#include <common.h> #include <dm.h> +#include <dm/device-internal.h> +#include <dm/lists.h> +#include <dm/uclass-internal.h> #include <dt-bindings/gpio/gpio.h> #include <errno.h> #include <fdtdec.h> @@ -141,6 +144,118 @@ static int gpio_find_and_xlate(struct gpio_desc *desc, return gpio_xlate_offs_flags(desc->dev, desc, args); }

+#if defined(CONFIG_DM_GPIO_HOG)

+struct gpio_hog_priv {

•   struct gpio_desc gpiod;
  

As above I don't think you need this.

Hmm.. when probing I need to request the gpio with gpio_dev_request_index() for which I need to pass a pointer to struct gpio_desc ...

The caller can declare this in a local variable.

The caller is always the device probe function, as gpio_hog_probe_all() loops over all gpio hog devices and probes them with device_probe() which cannot return a gpio_desc pointer, nor can such a pointer passed to.

Why isn't it valid to store in each gpio hog device the gpio desc it belongs too in device private data?

I don't see any point. Your caller needs to store the pointer anyway, or if not, could just as easily have a local variable with the info.

Also, passing a pointer to the struct is how we consistently do it so far.

What is the benefit of storing it in the callee rather than the caller?

...

...

And later I want to have the possibility to access the gpio for example with dm_gpio_set_value() and need therefore the gpio descriptor, so I must store it somewhere. Or do I overlook something obvious?

In that case the caller should store it, perhaps in a local variable, or perhaps in a private data structure if needed for a long time.

[..]

Of course I can introduce a list for each gpio hog definition filled when calling gpio_hog_probe_all() ... but than I have a second list which uses memory, may has to keep in sync the list if device are removed ... I want to avoid such an approach ... as we have already a list of gpio hog devices and the priv storage pointer for device private data.

I understand what you are saying about devices being removed, but actually removable is fine, it's only unbinding that is dangerous. And we already have that problem. We would need to add refcounts to fix it, and given that bootloaders don't normally need to unbind, I have not worried about it too much.

The approach I suggest would not use any more memory I think.

What do you think?

Well I still think it is better to have the caller own the struct.

...

...

...

...

•   return 0;
  

+}

+int gpio_hog_probe_all(void) +{

•   struct udevice *dev;
  

•   int ret;
  

•   for (uclass_first_device(UCLASS_NOP, &dev);
  

Perhaps you should create a new uclass for this?

Hmm.. yes, I had an approach with a new uclass, but I thought, the new [1] UCLASS_NOP is valid for this here to?

[1] http://patchwork.ozlabs.org/patch/1098913/

I see your point and I can see arguments for each side. But the fact that you have to check the driver suggests to me that a new uclass is better.

[..]

Ok with this approach the gpio_hog additions to gpio-uclass can be moved to a gpio-hog-uclass.c file. Also the problem where to store the gpio descriptor moves from device private data to uclass private data?

I don't think you have explained why you actually need to store the GPIO descriptor. Where do you use it later? Is it just the gpio_hog_probe_all() function?

Is it worth?

...

...

...

...

+/**

• • gpio_hog_probe_all() - probe all gpio devices with
• • gpio-hog subnodes.
• • @return: Returns return value from device_probe()

What if one fails and the others succeed?

Currently I return with error code from device_probe(). So the rest of the gpio-hogs are not probed.

Hmm.. may it is better to emmit a debug message and continue?

Sounds reasonable. But it seems to me that if there is an error there should be some visible report, so this could perhaps return the error code? Then the caller could print a messag?

Ok, so I add a printf message. I like more to add it in gpio_hog_probe_all(), because more than one gpio_hog can fail.

Regards, Simon