gpio: add DT support for basic memory-mapped GPIOs

This patch series adds device tree support for basic memory-mapped GPIOs. The GPIO library already allows drivers and architecture support code to reuse generic code for manageing a GPIO chip. Currently, a developer has to create a platform device "basic-mmio-gpio" and attach a bgpio_pdata platform data structure to make use of it. However, for architectures which rely on the device tree to enumerate devices, creating custom platform devices is another extra step that can be avoided by having direct support via a device tree binding.

I initially came across this patch [0] from Álvaro Fernández Rojas, while looking for an easy way to add support for the GPIO of my WD MyBook Live [1] (APM82181 - ppc464). This generic approach patch allowed me to easily get the GPIO (and the connected LEDs, buttons, gpiohogs, etc.) up and running. Even tought, Mr. Fernandez initially developed it for his work on the brcm63xx [2].

[0] <https://patchwork.ozlabs.org/patch/422121/>
[1] <https://github.com/chunkeey/MBL-openwrt>
[2] <https://wiki.openwrt.org/doc/hardware/soc/soc.broadcom.bcm63xx>

---
Note: Since it was such a long time ago, I had issues with getting all the replies from lkml.org (the side simply browned-out some replies and comments to the original, v2 and v3 series. So please if you have already had a comment on the previous patch that wasn't implemented in this RFC, let me know.

Thanks! (Please keep me in the CC)
---

Christian Lamparter (1):
gpio: generic: fix GPIO_GENERIC_PLATFORM is set to module case

Álvaro Fernández Rojas (2):
gpio: dt-bindings: add basic-mmio-gpio bindings
gpio: generic: add DT support for basic memory-mapped GPIOs

 .../devicetree/bindings/gpio/basic-mmio-gpio.txt   | 73 ++++++++++++++++++++++
 drivers/gpio/gpio-generic.c                        | 72 ++++++++++++++++++++-
 2 files changed, 142 insertions(+), 3 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/gpio/basic-mmio-gpio.txt

2.8.1

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

From: Álvaro Fernández Rojas <[email protected]>

This patch adds the device tree bindings for the basic-mmio-gpio. The basic-mmio-gpio is already part of a the GPIO generic library and shares its compatible with the platform device.

---
 .../devicetree/bindings/gpio/basic-mmio-gpio.txt   | 73 ++++++++++++++++++++++
 1 file changed, 73 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/gpio/basic-mmio-gpio.txt

new file mode 100644
index 0000000..5efb155
@@ -0,0 +1,73 @@
+Bindings for the generic driver for memory-mapped GPIO controllers.
+
+Required properties:
+	- compatible: should be "basic-mmio-gpio"
+	- reg-names: must contain
+		"dat" - data register
+		may contain
+		"set" - data set register
+		"clr" - data clear register
+		"dirout" - direction output register
+		"dirin" - direction input register
+	- reg: address + size pairs describing the GPIO register sets;
+		order must correspond with the order of entries in reg-names
+	- #gpio-cells = must be set to 2
+	- gpio-controller: Marks the device node as a gpio controller.
+
+Optional properties:
+	- ngpio: specifies the number of gpio mapped in the register.
+	- big-endian: force big endian register accesses.
+	- big-endian-byte-order: assign GPIOs in reverse order.
+	- unreadable-reg-set: data set register is not readable.
+	- read-output-reg-set: cache value set for reads.
+	- unreadable-reg-dir: dirout/dirin register is not readable.
+	- no-output: GPIOs are read-only.
+
+The GPIO generic library provides support for basic memory-mapped GPIO
+controllers. The configuration is detected by which resources are present.
+The simplest form of a GPIO controller that the driver support is just a
+single "dat" register, where GPIO state can be read and/or written.
+However, the driver supports far more:
+	- 8/16/32/64 bits registers. The number of GPIOs is automatically
+	  determined by the width of the registers.
+	- GPIO controllers with clear/set registers.
+	- GPIO controllers with a single "dat" register.
+	- Big endian bits/GPIOs ordering.
+
+For setting GPIO's there are three configurations:
+	1. single input/output register resource (named "dat"),
+	2. set/clear pair (named "set" and "clr"),
+	3. single output register resource and single input resource
+	   ("set" and dat").
+
+For setting the GPIO direction, there are three configurations:
+	a. simple bidirection GPIO that requires no configuration.
+	b. an output direction register (named "dirout")
+	   where a 1 bit indicates the GPIO is an output.
+	c. an input direction register (named "dirin")
+	   where a 1 bit indicates the GPIO is an input.
+
+Examples:
+
+	/* Configuration for single input/output register
+	 * for eight simple bidirection GPIOs.
+	 */
+	gpio_a_1 {
+		compatible = "basic-mmio-gpio";
+		reg = <0x18000000 0x1>;
+		reg-names = "dat";
+		#gpio-cells = <2>;
+		gpio-controller;
+	};
+
+	/* Configuration for set/clear pair registers with
+	 * 32 output direction register GPIOs.
+	 */
+	gpio_b_2 {
+		compatible = "basic-mmio-gpio";
+		reg = <0x18000000 0x4>, <0x18000010 0x4>,
+		      <0x18000004 0x4>, <0x18000008 0x4>;
+		reg-names = "dat", "set", "clr", "dirout";
+		#gpio-cells = <2>;
+		gpio-controller;
+	};

2.8.1

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

These things always start out "simple", "basic" or "generic". Then people extend them a property or 2 at a time until they are no longer basic. Make a bunch of GPIO drivers use a generic driver then maybe I'll be convinced this is a good idea. The concepts of generic code and generic bindings don't have to be coupled.

Rob
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

Well, I mentioned brcm63xx (MIPS) and MyBook Live (PPC). But OK, if you want me to look at the existing code in the kernel. I found the following candidates.

1. MOXA ART: "moxa,moxart-gpio"
http://lxr.free-electrons.com/source/drivers/gpio/gpio-moxart.c
This driver is used by the:
arch/arm/boot/dts/moxart.dtsi

---

index 1fd27ed..26707d0 100644
@@ -66,8 +66,11 @@
 		gpio: gpio@98700000 {
 			gpio-controller;
 			#gpio-cells = <2>;
-			compatible = "moxa,moxart-gpio";
-			reg = <0x98700000 0xC>;
+			compatible = "basic-mmio-gpio";
+			reg = <0x98700000 0x4
+			       0x98700004 0x4
+			       0x98700008 0x4>;
+			reg-names = "set dat dirout";
 		};
 
 		rtc: rtc {

---

2. GE FPGA based GPIO: "ge,imp3a-gpio", "gef,sbc310-gpio", "gef,sbc610-gpio"
http://lxr.free-electrons.com/source/drivers/gpio/gpio-ge.c

This driver is used by the:
arch/powerpc/boot/dts/fsl/ge_imp3a.dts
arch/powerpc/boot/dts/fsl/gef_ppc9a.dts
arch/powerpc/boot/dts/fsl/gef_sbc610.dts
arch/powerpc/boot/dts/fsl/gef_sbc310.dts

---

index 0424fc2..20dd084 100644
@@ -112,9 +112,14 @@
 		};
 		gef_gpio: gpio@7,14000 {
 			#gpio-cells = <2>;
-			compatible = "gef,ppc9a-gpio", "gef,sbc610-gpio";
-			reg = <0x7 0x14000 0x24>;
+			compatible = "basic-mmio-gpio";
+			reg = <0x7 0x14000 0x4
+			       0x7 0x14004 0x4
+			       0x7 0x14008 0x4>;
+			reg-names = "dirout dat set";
+			ngpio = <19>;
 			gpio-controller;
+			big-endian;
 		};
 	};

index a2bb47f..ac5855a 100644
@@ -93,9 +93,13 @@
 
 		gef_gpio: gpio@4,400 {
 			#gpio-cells = <2>;
-			compatible = "ge,imp3a-gpio";
-			reg = <0x4 0x400 0x24>;
+			compatible = "basic-mmio-gpio";
+			reg = <0x4 0x400 0x02
+			       0x4 0x404 0x02
+			       0x4 0x408 0x02>;
+			reg-names = "dirin dat set";
 			gpio-controller;
+			big-endian;
 		};
 
 		wdt@4,800 {

index 84b3d38..f09528f 100644
@@ -113,9 +113,14 @@
 		};
 		gef_gpio: gpio@4,8000 {
 			#gpio-cells = <2>;
-			compatible = "gef,sbc310-gpio";
-			reg = <0x4 0x8000 0x24>;
+			compatible = "basic-mmio-gpio";
+			reg = <0x4 0x8000 0x4
+			       0x4 0x8004 0x4
+			       0x4 0x8008 0x4>;
+			reg-names = "dirout dat set";
+			ngpio = <6>;
 			gpio-controller;
+			big-endian;
 		};
 	};
 

index 974446a..a9d740a 100644
@@ -110,9 +110,14 @@
 		};
 		gef_gpio: gpio@7,14000 {
 			#gpio-cells = <2>;
-			compatible = "gef,sbc610-gpio";
-			reg = <0x7 0x14000 0x24>;
+			compatible = "basic-mmio-gpio";
+			reg = <0x7 0x14000 0x4
+			       0x7 0x14004 0x4
+			       0x7 0x14008 0x4>;
+			reg-names = "dirout dat set";
+			ngpio = <19>;
 			gpio-controller;
+			big-endian;
 		};
 	};
---

3. CLPS711X GPIO driver: "cirrus,clps711x-gpio"
http://lxr.free-electrons.com/source/drivers/gpio/gpio-clps711x.c
(But there is no DTS which uses the in the kernel)

4. GPIO driver for the TS-4800 board: "technologic,ts4800-gpio" (Again no DTS in kernel)

What do you think. There's a total of six devices that can make use of the "basic-mmio-gpio". If this gives me your go-ahead, I'll include them in the series as proper patches.

Regards,
Christian
--
To unsubscribe from this list: send the line "unsubscribe linux-gpio" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

What I meant was convert the drivers, not the dts files to a generic driver. There is no benefit here from a dts perspective. It is driver consolidation that would make all this worthwhile.

Those are a good start. My other worry about this is next someone will want to add interrupt registers which is probably a majority of GPIO controllers.

Of course, I'm sure Linus has an opinion on all of this.

Rob
--
To unsubscribe from this list: send the line "unsubscribe linux-gpio" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

I'm not Linus, but after a quick look at the drivers mentioned by Christian I see little room for improvement (but still some - see below). Like Rob mentioned we want the DT's compatible property to describe the hardware as precisely as possible. "basic-mmio-gpio" is not precise and not backward-compatible. The examples listed by Christian are already all using the generic MMIO driver and are thus already quite short.

However, they all repeat the same basic sequence to initialize the generic IOMMU driver, and maybe this could be consolidated in a single file, with each matching chip having the arguments passed to bgpio_init() encoded as data (a bit like the simple-panel driver does
- see the platform_of_match member of drivers/gpu/drm/panel/panel-simple.c)

This would reduce the amount of code that basically does the same thing, and remove a few files from drivers/gpio. Even there I see that gpio-clps711x.c has some specifics compared to the other drivers, but maybe Christian can iron them out.

So without touching the DT I think consolidating the bgpio users in a fashion similar to what simple-panel does would be worthwhile. --
To unsubscribe from this list: send the line "unsubscribe linux-gpio" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

I'm getting ready to update my series with a patch like that. I would like to consolidate the gpio drivers into a single file. But for the v4 series at least the parsers are still residing in their original files (I hoped this would make the transition from the _probe to the parser less painless but the diff came out pretty much unreadable). I think this would be just a matter of what's preferred.

From: Álvaro Fernández Rojas <[email protected]>

This patch adds support for defining basic memory-mapped GPIOs within the device tree framework.

---
 drivers/gpio/gpio-generic.c | 70 +++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 68 insertions(+), 2 deletions(-)

index 54cddfa..f535299 100644
@@ -61,6 +61,8 @@ o        `                     ~~~~\___/~~~~    ` controller in FPGA is ,.`
 #include <linux/bitops.h>
 #include <linux/platform_device.h>
 #include <linux/mod_devicetable.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
 
 static void bgpio_write8(void __iomem *reg, unsigned long data)
 {

@@ -569,6 +571,59 @@ static void __iomem *bgpio_map(struct platform_device *pdev,
 	return devm_ioremap_resource(&pdev->dev, r);
 }
 
+#ifdef CONFIG_OF
+static const struct of_device_id  bgpio_of_match[] = {
+	{ .compatible = "basic-mmio-gpio" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, bgpio_of_match);
+
+static struct bgpio_pdata *bgpio_parse_dt(struct device *dev,
+					  unsigned long *flags)
+{
+	struct bgpio_pdata *pdata;
+	int err;
+
+	if (!of_match_node(bgpio_of_match, dev->of_node))
+		return NULL;
+
+	pdata = devm_kzalloc(dev, sizeof(struct bgpio_pdata),
+			     GFP_KERNEL);
+	if (!pdata)
+		return ERR_PTR(-ENOMEM);
+
+	err = of_property_read_u32(dev->of_node, "ngpio", &pdata->ngpio);
+	if (err && err != -EINVAL)
+		return ERR_PTR(err);
+
+	if (of_device_is_big_endian(dev->of_node))
+		*flags |= BGPIOF_BIG_ENDIAN_BYTE_ORDER;
+
+	if (of_property_read_bool(dev->of_node, "unreadable-reg-set"))
+		*flags |= BGPIOF_UNREADABLE_REG_SET;
+
+	if (of_property_read_bool(dev->of_node, "unreadable-reg-dir"))
+		*flags |= BGPIOF_UNREADABLE_REG_DIR;
+
+	if (of_property_read_bool(dev->of_node, "big-endian-byte-order"))
+		*flags |= BGPIOF_BIG_ENDIAN;
+
+	if (of_property_read_bool(dev->of_node, "read-output-reg-set"))
+		*flags |= BGPIOF_READ_OUTPUT_REG_SET;
+
+	if (of_property_read_bool(dev->of_node, "no-output"))
+		*flags |= BGPIOF_NO_OUTPUT;
+
+	return pdata;
+}
+#else
+static struct bgpio_data *bgpio_parse_dt(struct device *dev,
+					 unsigned long *flags)
+{
+	return NULL;
+}
+#endif /* CONFIG_OF */
+
 static int bgpio_pdev_probe(struct platform_device *pdev)
 {
 	struct device *dev = &pdev->dev;

@@ -579,10 +634,18 @@ static int bgpio_pdev_probe(struct platform_device *pdev)
 	void __iomem *dirout;
 	void __iomem *dirin;
 	unsigned long sz;
-	unsigned long flags = pdev->id_entry->driver_data;
+	unsigned long flags = 0;
 	int err;
 	struct gpio_chip *gc;
-	struct bgpio_pdata *pdata = dev_get_platdata(dev);
+	struct bgpio_pdata *pdata;
+
+	pdata = bgpio_parse_dt(dev, &flags);
+	if (IS_ERR(pdata)) {
+		return PTR_ERR(pdata);
+	} else if (!pdata) {
+		pdata = dev_get_platdata(dev);
+		flags = pdev->id_entry->driver_data;
+	}
 
 	r = platform_get_resource_byname(pdev, IORESOURCE_MEM, "dat");
 	if (!r)

@@ -646,6 +709,9 @@ MODULE_DEVICE_TABLE(platform, bgpio_id_table);
 static struct platform_driver bgpio_driver = {
 	.driver = {
 		.name = "basic-mmio-gpio",
+#ifdef CONFIG_OF
+		.of_match_table = of_match_ptr(bgpio_of_match),
+#endif
 	},
 	.id_table = bgpio_id_table,
 	.probe = bgpio_pdev_probe,

2.8.1

--
To unsubscribe from this list: send the line "unsubscribe linux-gpio" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

From: Álvaro Fernández Rojas <[email protected]>

This patch adds support for defining basic memory-mapped GPIOs within the device tree framework.

---
fixed bgpio_>>P<<data typo.
---

 drivers/gpio/gpio-generic.c | 70 +++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 68 insertions(+), 2 deletions(-)

index 54cddfa..f535299 100644
@@ -61,6 +61,8 @@ o        `                     ~~~~\___/~~~~    ` controller in FPGA is ,.`
 #include <linux/bitops.h>
 #include <linux/platform_device.h>
 #include <linux/mod_devicetable.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
 
 static void bgpio_write8(void __iomem *reg, unsigned long data)
 {

@@ -569,6 +571,59 @@ static void __iomem *bgpio_map(struct platform_device *pdev,
 	return devm_ioremap_resource(&pdev->dev, r);
 }
 
+#ifdef CONFIG_OF
+static const struct of_device_id  bgpio_of_match[] = {
+	{ .compatible = "basic-mmio-gpio" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, bgpio_of_match);
+
+static struct bgpio_pdata *bgpio_parse_dt(struct device *dev,
+					  unsigned long *flags)
+{
+	struct bgpio_pdata *pdata;
+	int err;
+
+	if (!of_match_node(bgpio_of_match, dev->of_node))
+		return NULL;
+
+	pdata = devm_kzalloc(dev, sizeof(struct bgpio_pdata),
+			     GFP_KERNEL);
+	if (!pdata)
+		return ERR_PTR(-ENOMEM);
+
+	err = of_property_read_u32(dev->of_node, "ngpio", &pdata->ngpio);
+	if (err && err != -EINVAL)
+		return ERR_PTR(err);
+
+	if (of_device_is_big_endian(dev->of_node))
+		*flags |= BGPIOF_BIG_ENDIAN_BYTE_ORDER;
+
+	if (of_property_read_bool(dev->of_node, "unreadable-reg-set"))
+		*flags |= BGPIOF_UNREADABLE_REG_SET;
+
+	if (of_property_read_bool(dev->of_node, "unreadable-reg-dir"))
+		*flags |= BGPIOF_UNREADABLE_REG_DIR;
+
+	if (of_property_read_bool(dev->of_node, "big-endian-byte-order"))
+		*flags |= BGPIOF_BIG_ENDIAN;
+
+	if (of_property_read_bool(dev->of_node, "read-output-reg-set"))
+		*flags |= BGPIOF_READ_OUTPUT_REG_SET;
+
+	if (of_property_read_bool(dev->of_node, "no-output"))
+		*flags |= BGPIOF_NO_OUTPUT;
+
+	return pdata;
+}
+#else
+static struct bgpio_pdata *bgpio_parse_dt(struct device *dev,
+					  unsigned long *flags)
+{
+	return NULL;
+}
+#endif /* CONFIG_OF */
+
 static int bgpio_pdev_probe(struct platform_device *pdev)
 {
 	struct device *dev = &pdev->dev;

@@ -579,10 +634,18 @@ static int bgpio_pdev_probe(struct platform_device *pdev)
 	void __iomem *dirout;
 	void __iomem *dirin;
 	unsigned long sz;
-	unsigned long flags = pdev->id_entry->driver_data;
+	unsigned long flags = 0;
 	int err;
 	struct gpio_chip *gc;
-	struct bgpio_pdata *pdata = dev_get_platdata(dev);
+	struct bgpio_pdata *pdata;
+
+	pdata = bgpio_parse_dt(dev, &flags);
+	if (IS_ERR(pdata)) {
+		return PTR_ERR(pdata);
+	} else if (!pdata) {
+		pdata = dev_get_platdata(dev);
+		flags = pdev->id_entry->driver_data;
+	}
 
 	r = platform_get_resource_byname(pdev, IORESOURCE_MEM, "dat");
 	if (!r)

@@ -646,6 +709,9 @@ MODULE_DEVICE_TABLE(platform, bgpio_id_table);
 static struct platform_driver bgpio_driver = {
 	.driver = {
 		.name = "basic-mmio-gpio",
+#ifdef CONFIG_OF
+		.of_match_table = of_match_ptr(bgpio_of_match),
+#endif
 	},
 	.id_table = bgpio_id_table,
 	.probe = bgpio_pdev_probe,

2.8.1

--
To unsubscribe from this list: send the line "unsubscribe linux-gpio" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

GPIO_GENERIC_PLATFORM is a tristate. If the module option is selected the resulting gpio-generic.ko will lack most of the module initialzation and probe code.

---
 drivers/gpio/gpio-generic.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index f535299..b2170d0 100644
@@ -551,7 +551,7 @@ int bgpio_init(struct gpio_chip *gc, struct device *dev,
 }
 EXPORT_SYMBOL_GPL(bgpio_init);
 
-#ifdef CONFIG_GPIO_GENERIC_PLATFORM
+#if IS_ENABLED(CONFIG_GPIO_GENERIC_PLATFORM)
 
 static void __iomem *bgpio_map(struct platform_device *pdev,
 			       const char *name,

2.8.1