All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH v5 0/3] Enable X-Gene standby GPIO as interrupt controller
@ 2016-01-29  4:28 ` Quan Nguyen
  0 siblings, 0 replies; 24+ messages in thread
From: Quan Nguyen @ 2016-01-29  4:28 UTC (permalink / raw)
  To: linus.walleij, linux-gpio, devicetree, linux-arm-kernel,
	Thomas Gleixner, Jason Cooper, Marc Zyngier
  Cc: Y Vo, Phong Vo, Loc Ho, Feng Kan, Duc Dang, patches, Quan Nguyen

V5 Changes:
        - Add optional properties to describe HW specifics
        - Use irq_create_fwspec_mapping() instead of low level functions
        - Move gpiochip_lock_as_irq() to activate method
        - Remove redundant checks
	- Add optional HW properties description to DT binding document
        - Update "interrupts" property description in DT binding document.
V4 Changes:
        - Convert to hierarchical irq domain
V3 Changes:
        - Picking up from version 2 of "Y Vo <yvo@apm.com>"
        - Get HW resource information from DT
        - Avoid keeping parent IRQ controller
V2 Changes:
        - Support X-Gene standby GPIO as an interrupt controller.

Quan Nguyen (3):
  gpio: xgene: Enable X-Gene standby GPIO as interrupt controller
  Documentation: gpio: Update description for X-Gene standby GPIO
    controller DTS binding
  arm64: dts: Update X-Gene standby GPIO controller DTS entries

 .../devicetree/bindings/gpio/gpio-xgene-sb.txt     |  47 +++-
 arch/arm64/boot/dts/apm/apm-storm.dtsi             |   3 +
 drivers/gpio/gpio-xgene-sb.c                       | 275 ++++++++++++++++++---
 3 files changed, 286 insertions(+), 39 deletions(-)

-- 
1.7.12.4


^ permalink raw reply	[flat|nested] 24+ messages in thread

* [PATCH v5 0/3] Enable X-Gene standby GPIO as interrupt controller
@ 2016-01-29  4:28 ` Quan Nguyen
  0 siblings, 0 replies; 24+ messages in thread
From: Quan Nguyen @ 2016-01-29  4:28 UTC (permalink / raw)
  To: linux-arm-kernel

V5 Changes:
        - Add optional properties to describe HW specifics
        - Use irq_create_fwspec_mapping() instead of low level functions
        - Move gpiochip_lock_as_irq() to activate method
        - Remove redundant checks
	- Add optional HW properties description to DT binding document
        - Update "interrupts" property description in DT binding document.
V4 Changes:
        - Convert to hierarchical irq domain
V3 Changes:
        - Picking up from version 2 of "Y Vo <yvo@apm.com>"
        - Get HW resource information from DT
        - Avoid keeping parent IRQ controller
V2 Changes:
        - Support X-Gene standby GPIO as an interrupt controller.

Quan Nguyen (3):
  gpio: xgene: Enable X-Gene standby GPIO as interrupt controller
  Documentation: gpio: Update description for X-Gene standby GPIO
    controller DTS binding
  arm64: dts: Update X-Gene standby GPIO controller DTS entries

 .../devicetree/bindings/gpio/gpio-xgene-sb.txt     |  47 +++-
 arch/arm64/boot/dts/apm/apm-storm.dtsi             |   3 +
 drivers/gpio/gpio-xgene-sb.c                       | 275 ++++++++++++++++++---
 3 files changed, 286 insertions(+), 39 deletions(-)

-- 
1.7.12.4

^ permalink raw reply	[flat|nested] 24+ messages in thread

* [PATCH v5 1/3] gpio: xgene: Enable X-Gene standby GPIO as interrupt controller
  2016-01-29  4:28 ` Quan Nguyen
@ 2016-01-29  4:28   ` Quan Nguyen
  -1 siblings, 0 replies; 24+ messages in thread
From: Quan Nguyen @ 2016-01-29  4:28 UTC (permalink / raw)
  To: linus.walleij, linux-gpio, devicetree, linux-arm-kernel,
	Thomas Gleixner, Jason Cooper, Marc Zyngier
  Cc: Y Vo, Phong Vo, Loc Ho, Feng Kan, Duc Dang, patches, Quan Nguyen

Enable X-Gene standby GPIO controller as interrupt controller to provide
its own resources. This avoids ambiguity where GIC interrupt resource is
use as X-Gene standby GPIO interrupt resource in user driver.

Signed-off-by: Y Vo <yvo@apm.com>
Signed-off-by: Quan Nguyen <qnguyen@apm.com>
---
 drivers/gpio/gpio-xgene-sb.c | 275 ++++++++++++++++++++++++++++++++++++++-----
 1 file changed, 243 insertions(+), 32 deletions(-)

diff --git a/drivers/gpio/gpio-xgene-sb.c b/drivers/gpio/gpio-xgene-sb.c
index 282004d..026da90 100644
--- a/drivers/gpio/gpio-xgene-sb.c
+++ b/drivers/gpio/gpio-xgene-sb.c
@@ -2,8 +2,9 @@
  * AppliedMicro X-Gene SoC GPIO-Standby Driver
  *
  * Copyright (c) 2014, Applied Micro Circuits Corporation
- * Author: 	Tin Huynh <tnhuynh@apm.com>.
- * 		Y Vo <yvo@apm.com>.
+ * Author:	Tin Huynh <tnhuynh@apm.com>.
+ *		Y Vo <yvo@apm.com>.
+ *		Quan Nguyen <qnguyen@apm.com>.
  *
  * This program is free software; you can redistribute  it and/or modify it
  * under  the terms of  the GNU General  Public License as published by the
@@ -28,9 +29,14 @@
 
 #include "gpiolib.h"
 
-#define XGENE_MAX_GPIO_DS		22
-#define XGENE_MAX_GPIO_DS_IRQ		6
+/* Common property names */
+#define XGENE_NIRQ_PROPERTY		"apm,nr-irqs"
+#define XGENE_NGPIO_PROPERTY		"apm,nr-gpios"
+#define XGENE_IRQ_START_PROPERTY	"apm,irq-start"
 
+#define XGENE_DFLT_MAX_NGPIO		22
+#define XGENE_DFLT_MAX_NIRQ		6
+#define XGENE_DFLT_IRQ_START_PIN	8
 #define GPIO_MASK(x)			(1U << ((x) % 32))
 
 #define MPA_GPIO_INT_LVL		0x0290
@@ -39,19 +45,32 @@
 #define MPA_GPIO_IN_ADDR 		0x02a4
 #define MPA_GPIO_SEL_LO 		0x0294
 
+#define GPIO_INT_LEVEL_H	0x000001
+#define GPIO_INT_LEVEL_L	0x000000
+
 /**
  * struct xgene_gpio_sb - GPIO-Standby private data structure.
  * @gc:				memory-mapped GPIO controllers.
- * @irq:			Mapping GPIO pins and interrupt number
- * nirq:			Number of GPIO pins that supports interrupt
+ * @regs:			GPIO register base offset
+ * @irq_domain:			GPIO interrupt domain
+ * @irq_start:			GPIO pin that start support interrupt
+ * @nirq:			Number of GPIO pins that supports interrupt
+ * @parent_irq_base:		Start parent HWIRQ
  */
 struct xgene_gpio_sb {
 	struct gpio_chip	gc;
-	u32 *irq;
-	u32 nirq;
+	void __iomem		*regs;
+	struct irq_domain	*irq_domain;
+	u16			irq_start;
+	u16			nirq;
+	u16			parent_irq_base;
 };
 
-static void xgene_gpio_set_bit(struct gpio_chip *gc, void __iomem *reg, u32 gpio, int val)
+#define HWIRQ_TO_GPIO(priv, hwirq) ((hwirq) + (priv)->irq_start)
+#define GPIO_TO_HWIRQ(priv, gpio) ((gpio) - (priv)->irq_start)
+
+static void xgene_gpio_set_bit(struct gpio_chip *gc,
+				void __iomem *reg, u32 gpio, int val)
 {
 	u32 data;
 
@@ -63,23 +82,180 @@ static void xgene_gpio_set_bit(struct gpio_chip *gc, void __iomem *reg, u32 gpio
 	gc->write_reg(reg, data);
 }
 
-static int apm_gpio_sb_to_irq(struct gpio_chip *gc, u32 gpio)
+static int xgene_gpio_sb_irq_set_type(struct irq_data *d, unsigned int type)
+{
+	struct xgene_gpio_sb *priv = irq_data_get_irq_chip_data(d);
+	int gpio = HWIRQ_TO_GPIO(priv, d->hwirq);
+	int lvl_type = GPIO_INT_LEVEL_H;
+
+	switch (type & IRQ_TYPE_SENSE_MASK) {
+	case IRQ_TYPE_EDGE_RISING:
+	case IRQ_TYPE_LEVEL_HIGH:
+		lvl_type = GPIO_INT_LEVEL_H;
+		break;
+	case IRQ_TYPE_EDGE_FALLING:
+	case IRQ_TYPE_LEVEL_LOW:
+		lvl_type = GPIO_INT_LEVEL_L;
+		break;
+	default:
+		break;
+	}
+
+	xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_SEL_LO,
+			gpio * 2, 1);
+	xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_INT_LVL,
+			d->hwirq, lvl_type);
+
+	/* Propagate IRQ type setting to parent */
+	if (type & IRQ_TYPE_EDGE_BOTH)
+		return irq_chip_set_type_parent(d, IRQ_TYPE_EDGE_RISING);
+	else
+		return irq_chip_set_type_parent(d, IRQ_TYPE_LEVEL_HIGH);
+}
+
+static void xgene_gpio_sb_irq_shutdown(struct irq_data *d)
+{
+	struct xgene_gpio_sb *priv = irq_data_get_irq_chip_data(d);
+
+	gpiochip_unlock_as_irq(&priv->gc, HWIRQ_TO_GPIO(priv, d->hwirq));
+}
+
+static struct irq_chip xgene_gpio_sb_irq_chip = {
+	.name           = "sbgpio",
+	.irq_ack        = irq_chip_ack_parent,
+	.irq_eoi	= irq_chip_eoi_parent,
+	.irq_mask       = irq_chip_mask_parent,
+	.irq_unmask     = irq_chip_unmask_parent,
+	.irq_set_type   = xgene_gpio_sb_irq_set_type,
+	.irq_shutdown   = xgene_gpio_sb_irq_shutdown,
+};
+
+static int xgene_gpio_sb_to_irq(struct gpio_chip *gc, u32 gpio)
 {
 	struct xgene_gpio_sb *priv = gpiochip_get_data(gc);
+	struct irq_fwspec fwspec;
+
+	if ((gpio < priv->irq_start) ||
+			(gpio > HWIRQ_TO_GPIO(priv, priv->nirq)))
+		return -ENXIO;
+
+	if (gc->parent->of_node)
+		fwspec.fwnode = of_node_to_fwnode(gc->parent->of_node);
+	else
+		fwspec.fwnode = gc->parent->fwnode;
+	fwspec.param_count = 2;
+	fwspec.param[0] = GPIO_TO_HWIRQ(priv, gpio);
+	fwspec.param[1] = IRQ_TYPE_NONE;
+	return irq_create_fwspec_mapping(&fwspec);
+}
+
+static void xgene_gpio_sb_domain_activate(struct irq_domain *d,
+		struct irq_data *irq_data)
+{
+	struct xgene_gpio_sb *priv = d->host_data;
+	u32 gpio = HWIRQ_TO_GPIO(priv, irq_data->hwirq);
+
+	if (gpiochip_lock_as_irq(&priv->gc, gpio)) {
+		dev_err(priv->gc.parent,
+		"Unable to configure XGene GPIO standby pin %d as IRQ\n",
+				gpio);
+		return;
+	}
+
+	xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_SEL_LO,
+			gpio * 2, 1);
+}
+
+static void xgene_gpio_sb_domain_deactivate(struct irq_domain *d,
+		struct irq_data *irq_data)
+{
+	struct xgene_gpio_sb *priv = d->host_data;
+	u32 gpio = HWIRQ_TO_GPIO(priv, irq_data->hwirq);
+
+	gpiochip_unlock_as_irq(&priv->gc, gpio);
+	xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_SEL_LO,
+			gpio * 2, 0);
+}
 
-	if (priv->irq[gpio])
-		return priv->irq[gpio];
+static int xgene_gpio_sb_domain_translate(struct irq_domain *d,
+		struct irq_fwspec *fwspec,
+		unsigned long *hwirq,
+		unsigned int *type)
+{
+	struct xgene_gpio_sb *priv = d->host_data;
+
+	if ((fwspec->param_count != 2) ||
+		(fwspec->param[0] >= priv->nirq))
+		return -EINVAL;
+	*hwirq = fwspec->param[0];
+	*type = fwspec->param[1];
+	return 0;
+}
+
+static int xgene_gpio_sb_domain_alloc(struct irq_domain *domain,
+					unsigned int virq,
+					unsigned int nr_irqs, void *data)
+{
+	struct irq_fwspec *fwspec = data;
+	struct irq_fwspec parent_fwspec;
+	struct xgene_gpio_sb *priv = domain->host_data;
+	irq_hw_number_t hwirq;
+	unsigned int i;
+
+	hwirq = fwspec->param[0];
+	for (i = 0; i < nr_irqs; i++)
+		irq_domain_set_hwirq_and_chip(domain, virq + i, hwirq + i,
+				&xgene_gpio_sb_irq_chip, priv);
+
+	if (is_of_node(domain->parent->fwnode)) {
+		parent_fwspec.fwnode = domain->parent->fwnode;
+		parent_fwspec.param_count = 3;
+		parent_fwspec.param[0] = 0;/* SPI */
+		/* Skip SGIs and PPIs*/
+		parent_fwspec.param[1] = hwirq + priv->parent_irq_base - 32;
+		parent_fwspec.param[2] = fwspec->param[1];
+	} else if (is_fwnode_irqchip(domain->parent->fwnode)) {
+		parent_fwspec.fwnode = domain->parent->fwnode;
+		parent_fwspec.param_count = 2;
+		parent_fwspec.param[0] = hwirq + priv->parent_irq_base;
+		parent_fwspec.param[1] = fwspec->param[1];
+	} else
+		return -EINVAL;
 
-	return -ENXIO;
+	return irq_domain_alloc_irqs_parent(domain, virq, nr_irqs,
+			&parent_fwspec);
 }
 
+static void xgene_gpio_sb_domain_free(struct irq_domain *domain,
+		unsigned int virq,
+		unsigned int nr_irqs)
+{
+	struct irq_data *d;
+	unsigned int i;
+
+	for (i = 0; i < nr_irqs; i++) {
+		d = irq_domain_get_irq_data(domain, virq + i);
+		irq_domain_reset_irq_data(d);
+	}
+}
+
+static const struct irq_domain_ops xgene_gpio_sb_domain_ops = {
+	.translate      = xgene_gpio_sb_domain_translate,
+	.alloc          = xgene_gpio_sb_domain_alloc,
+	.free           = xgene_gpio_sb_domain_free,
+	.activate	= xgene_gpio_sb_domain_activate,
+	.deactivate	= xgene_gpio_sb_domain_deactivate,
+};
+
 static int xgene_gpio_sb_probe(struct platform_device *pdev)
 {
 	struct xgene_gpio_sb *priv;
-	u32 ret, i;
-	u32 default_lines[] = {0x08, 0x09, 0x0A, 0x0B, 0x0C, 0x0D};
+	u32 ret;
 	struct resource *res;
 	void __iomem *regs;
+	struct irq_domain *parent_domain = NULL;
+	struct fwnode_handle *fwnode;
+	u32 val32;
 
 	priv = devm_kzalloc(&pdev->dev, sizeof(*priv), GFP_KERNEL);
 	if (!priv)
@@ -90,6 +266,18 @@ static int xgene_gpio_sb_probe(struct platform_device *pdev)
 	if (IS_ERR(regs))
 		return PTR_ERR(regs);
 
+	priv->regs = regs;
+
+	ret = platform_get_irq(pdev, 0);
+	if (ret > 0) {
+		priv->parent_irq_base = irq_get_irq_data(ret)->hwirq;
+		parent_domain = irq_get_irq_data(ret)->domain;
+	}
+	if (!parent_domain) {
+		dev_err(&pdev->dev, "unable to obtain parent domain\n");
+		return -ENODEV;
+	}
+
 	ret = bgpio_init(&priv->gc, &pdev->dev, 4,
 			regs + MPA_GPIO_IN_ADDR,
 			regs + MPA_GPIO_OUT_ADDR, NULL,
@@ -97,30 +285,51 @@ static int xgene_gpio_sb_probe(struct platform_device *pdev)
         if (ret)
                 return ret;
 
-	priv->gc.to_irq = apm_gpio_sb_to_irq;
-	priv->gc.ngpio = XGENE_MAX_GPIO_DS;
+	priv->gc.to_irq = xgene_gpio_sb_to_irq;
 
-	priv->nirq = XGENE_MAX_GPIO_DS_IRQ;
+	/* Retrieve start irq pin, use default if property not found */
+	priv->irq_start = XGENE_DFLT_IRQ_START_PIN;
+	if (!device_property_read_u32(&pdev->dev,
+					XGENE_IRQ_START_PROPERTY, &val32))
+		priv->irq_start = val32;
 
-	priv->irq = devm_kzalloc(&pdev->dev, sizeof(u32) * XGENE_MAX_GPIO_DS,
-				   GFP_KERNEL);
-	if (!priv->irq)
-		return -ENOMEM;
+	/* Retrieve number irqs, use default if property not found */
+	priv->nirq = XGENE_DFLT_MAX_NIRQ;
+	if (!device_property_read_u32(&pdev->dev, XGENE_NIRQ_PROPERTY, &val32))
+		priv->nirq = val32;
 
-	for (i = 0; i < priv->nirq; i++) {
-		priv->irq[default_lines[i]] = platform_get_irq(pdev, i);
-		xgene_gpio_set_bit(&priv->gc, regs + MPA_GPIO_SEL_LO,
-                                   default_lines[i] * 2, 1);
-		xgene_gpio_set_bit(&priv->gc, regs + MPA_GPIO_INT_LVL, i, 1);
-	}
+	/* Retrieve number gpio, use default if property not found */
+	priv->gc.ngpio = XGENE_DFLT_MAX_NGPIO;
+	if (!device_property_read_u32(&pdev->dev, XGENE_NGPIO_PROPERTY, &val32))
+		priv->gc.ngpio = val32;
+
+	dev_info(&pdev->dev, "Support %d gpios, %d irqs start from pin %d\n",
+			priv->gc.ngpio, priv->nirq, priv->irq_start);
 
 	platform_set_drvdata(pdev, priv);
 
-	ret = gpiochip_add_data(&priv->gc, priv);
-	if (ret)
-		dev_err(&pdev->dev, "failed to register X-Gene GPIO Standby driver\n");
+	if (pdev->dev.of_node)
+		fwnode = of_node_to_fwnode(pdev->dev.of_node);
 	else
-		dev_info(&pdev->dev, "X-Gene GPIO Standby driver registered\n");
+		fwnode = pdev->dev.fwnode;
+
+	priv->irq_domain = irq_domain_create_hierarchy(parent_domain,
+					0, priv->nirq, fwnode,
+					&xgene_gpio_sb_domain_ops, priv);
+	if (!priv->irq_domain)
+		return -ENODEV;
+
+	priv->gc.irqdomain = priv->irq_domain;
+
+	ret = gpiochip_add_data(&priv->gc, priv);
+	if (ret) {
+		dev_err(&pdev->dev,
+			"failed to register X-Gene GPIO Standby driver\n");
+		irq_domain_remove(priv->irq_domain);
+		return ret;
+	}
+
+	dev_info(&pdev->dev, "X-Gene GPIO Standby driver registered\n");
 
 	if (priv->nirq > 0) {
 		/* Register interrupt handlers for gpio signaled acpi events */
@@ -138,6 +347,8 @@ static int xgene_gpio_sb_remove(struct platform_device *pdev)
 		acpi_gpiochip_free_interrupts(&priv->gc);
 	}
 
+	irq_domain_remove(priv->irq_domain);
+
 	gpiochip_remove(&priv->gc);
 	return 0;
 }
-- 
1.7.12.4


^ permalink raw reply related	[flat|nested] 24+ messages in thread

* [PATCH v5 1/3] gpio: xgene: Enable X-Gene standby GPIO as interrupt controller
@ 2016-01-29  4:28   ` Quan Nguyen
  0 siblings, 0 replies; 24+ messages in thread
From: Quan Nguyen @ 2016-01-29  4:28 UTC (permalink / raw)
  To: linux-arm-kernel

Enable X-Gene standby GPIO controller as interrupt controller to provide
its own resources. This avoids ambiguity where GIC interrupt resource is
use as X-Gene standby GPIO interrupt resource in user driver.

Signed-off-by: Y Vo <yvo@apm.com>
Signed-off-by: Quan Nguyen <qnguyen@apm.com>
---
 drivers/gpio/gpio-xgene-sb.c | 275 ++++++++++++++++++++++++++++++++++++++-----
 1 file changed, 243 insertions(+), 32 deletions(-)

diff --git a/drivers/gpio/gpio-xgene-sb.c b/drivers/gpio/gpio-xgene-sb.c
index 282004d..026da90 100644
--- a/drivers/gpio/gpio-xgene-sb.c
+++ b/drivers/gpio/gpio-xgene-sb.c
@@ -2,8 +2,9 @@
  * AppliedMicro X-Gene SoC GPIO-Standby Driver
  *
  * Copyright (c) 2014, Applied Micro Circuits Corporation
- * Author: 	Tin Huynh <tnhuynh@apm.com>.
- * 		Y Vo <yvo@apm.com>.
+ * Author:	Tin Huynh <tnhuynh@apm.com>.
+ *		Y Vo <yvo@apm.com>.
+ *		Quan Nguyen <qnguyen@apm.com>.
  *
  * This program is free software; you can redistribute  it and/or modify it
  * under  the terms of  the GNU General  Public License as published by the
@@ -28,9 +29,14 @@
 
 #include "gpiolib.h"
 
-#define XGENE_MAX_GPIO_DS		22
-#define XGENE_MAX_GPIO_DS_IRQ		6
+/* Common property names */
+#define XGENE_NIRQ_PROPERTY		"apm,nr-irqs"
+#define XGENE_NGPIO_PROPERTY		"apm,nr-gpios"
+#define XGENE_IRQ_START_PROPERTY	"apm,irq-start"
 
+#define XGENE_DFLT_MAX_NGPIO		22
+#define XGENE_DFLT_MAX_NIRQ		6
+#define XGENE_DFLT_IRQ_START_PIN	8
 #define GPIO_MASK(x)			(1U << ((x) % 32))
 
 #define MPA_GPIO_INT_LVL		0x0290
@@ -39,19 +45,32 @@
 #define MPA_GPIO_IN_ADDR 		0x02a4
 #define MPA_GPIO_SEL_LO 		0x0294
 
+#define GPIO_INT_LEVEL_H	0x000001
+#define GPIO_INT_LEVEL_L	0x000000
+
 /**
  * struct xgene_gpio_sb - GPIO-Standby private data structure.
  * @gc:				memory-mapped GPIO controllers.
- * @irq:			Mapping GPIO pins and interrupt number
- * nirq:			Number of GPIO pins that supports interrupt
+ * @regs:			GPIO register base offset
+ * @irq_domain:			GPIO interrupt domain
+ * @irq_start:			GPIO pin that start support interrupt
+ * @nirq:			Number of GPIO pins that supports interrupt
+ * @parent_irq_base:		Start parent HWIRQ
  */
 struct xgene_gpio_sb {
 	struct gpio_chip	gc;
-	u32 *irq;
-	u32 nirq;
+	void __iomem		*regs;
+	struct irq_domain	*irq_domain;
+	u16			irq_start;
+	u16			nirq;
+	u16			parent_irq_base;
 };
 
-static void xgene_gpio_set_bit(struct gpio_chip *gc, void __iomem *reg, u32 gpio, int val)
+#define HWIRQ_TO_GPIO(priv, hwirq) ((hwirq) + (priv)->irq_start)
+#define GPIO_TO_HWIRQ(priv, gpio) ((gpio) - (priv)->irq_start)
+
+static void xgene_gpio_set_bit(struct gpio_chip *gc,
+				void __iomem *reg, u32 gpio, int val)
 {
 	u32 data;
 
@@ -63,23 +82,180 @@ static void xgene_gpio_set_bit(struct gpio_chip *gc, void __iomem *reg, u32 gpio
 	gc->write_reg(reg, data);
 }
 
-static int apm_gpio_sb_to_irq(struct gpio_chip *gc, u32 gpio)
+static int xgene_gpio_sb_irq_set_type(struct irq_data *d, unsigned int type)
+{
+	struct xgene_gpio_sb *priv = irq_data_get_irq_chip_data(d);
+	int gpio = HWIRQ_TO_GPIO(priv, d->hwirq);
+	int lvl_type = GPIO_INT_LEVEL_H;
+
+	switch (type & IRQ_TYPE_SENSE_MASK) {
+	case IRQ_TYPE_EDGE_RISING:
+	case IRQ_TYPE_LEVEL_HIGH:
+		lvl_type = GPIO_INT_LEVEL_H;
+		break;
+	case IRQ_TYPE_EDGE_FALLING:
+	case IRQ_TYPE_LEVEL_LOW:
+		lvl_type = GPIO_INT_LEVEL_L;
+		break;
+	default:
+		break;
+	}
+
+	xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_SEL_LO,
+			gpio * 2, 1);
+	xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_INT_LVL,
+			d->hwirq, lvl_type);
+
+	/* Propagate IRQ type setting to parent */
+	if (type & IRQ_TYPE_EDGE_BOTH)
+		return irq_chip_set_type_parent(d, IRQ_TYPE_EDGE_RISING);
+	else
+		return irq_chip_set_type_parent(d, IRQ_TYPE_LEVEL_HIGH);
+}
+
+static void xgene_gpio_sb_irq_shutdown(struct irq_data *d)
+{
+	struct xgene_gpio_sb *priv = irq_data_get_irq_chip_data(d);
+
+	gpiochip_unlock_as_irq(&priv->gc, HWIRQ_TO_GPIO(priv, d->hwirq));
+}
+
+static struct irq_chip xgene_gpio_sb_irq_chip = {
+	.name           = "sbgpio",
+	.irq_ack        = irq_chip_ack_parent,
+	.irq_eoi	= irq_chip_eoi_parent,
+	.irq_mask       = irq_chip_mask_parent,
+	.irq_unmask     = irq_chip_unmask_parent,
+	.irq_set_type   = xgene_gpio_sb_irq_set_type,
+	.irq_shutdown   = xgene_gpio_sb_irq_shutdown,
+};
+
+static int xgene_gpio_sb_to_irq(struct gpio_chip *gc, u32 gpio)
 {
 	struct xgene_gpio_sb *priv = gpiochip_get_data(gc);
+	struct irq_fwspec fwspec;
+
+	if ((gpio < priv->irq_start) ||
+			(gpio > HWIRQ_TO_GPIO(priv, priv->nirq)))
+		return -ENXIO;
+
+	if (gc->parent->of_node)
+		fwspec.fwnode = of_node_to_fwnode(gc->parent->of_node);
+	else
+		fwspec.fwnode = gc->parent->fwnode;
+	fwspec.param_count = 2;
+	fwspec.param[0] = GPIO_TO_HWIRQ(priv, gpio);
+	fwspec.param[1] = IRQ_TYPE_NONE;
+	return irq_create_fwspec_mapping(&fwspec);
+}
+
+static void xgene_gpio_sb_domain_activate(struct irq_domain *d,
+		struct irq_data *irq_data)
+{
+	struct xgene_gpio_sb *priv = d->host_data;
+	u32 gpio = HWIRQ_TO_GPIO(priv, irq_data->hwirq);
+
+	if (gpiochip_lock_as_irq(&priv->gc, gpio)) {
+		dev_err(priv->gc.parent,
+		"Unable to configure XGene GPIO standby pin %d as IRQ\n",
+				gpio);
+		return;
+	}
+
+	xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_SEL_LO,
+			gpio * 2, 1);
+}
+
+static void xgene_gpio_sb_domain_deactivate(struct irq_domain *d,
+		struct irq_data *irq_data)
+{
+	struct xgene_gpio_sb *priv = d->host_data;
+	u32 gpio = HWIRQ_TO_GPIO(priv, irq_data->hwirq);
+
+	gpiochip_unlock_as_irq(&priv->gc, gpio);
+	xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_SEL_LO,
+			gpio * 2, 0);
+}
 
-	if (priv->irq[gpio])
-		return priv->irq[gpio];
+static int xgene_gpio_sb_domain_translate(struct irq_domain *d,
+		struct irq_fwspec *fwspec,
+		unsigned long *hwirq,
+		unsigned int *type)
+{
+	struct xgene_gpio_sb *priv = d->host_data;
+
+	if ((fwspec->param_count != 2) ||
+		(fwspec->param[0] >= priv->nirq))
+		return -EINVAL;
+	*hwirq = fwspec->param[0];
+	*type = fwspec->param[1];
+	return 0;
+}
+
+static int xgene_gpio_sb_domain_alloc(struct irq_domain *domain,
+					unsigned int virq,
+					unsigned int nr_irqs, void *data)
+{
+	struct irq_fwspec *fwspec = data;
+	struct irq_fwspec parent_fwspec;
+	struct xgene_gpio_sb *priv = domain->host_data;
+	irq_hw_number_t hwirq;
+	unsigned int i;
+
+	hwirq = fwspec->param[0];
+	for (i = 0; i < nr_irqs; i++)
+		irq_domain_set_hwirq_and_chip(domain, virq + i, hwirq + i,
+				&xgene_gpio_sb_irq_chip, priv);
+
+	if (is_of_node(domain->parent->fwnode)) {
+		parent_fwspec.fwnode = domain->parent->fwnode;
+		parent_fwspec.param_count = 3;
+		parent_fwspec.param[0] = 0;/* SPI */
+		/* Skip SGIs and PPIs*/
+		parent_fwspec.param[1] = hwirq + priv->parent_irq_base - 32;
+		parent_fwspec.param[2] = fwspec->param[1];
+	} else if (is_fwnode_irqchip(domain->parent->fwnode)) {
+		parent_fwspec.fwnode = domain->parent->fwnode;
+		parent_fwspec.param_count = 2;
+		parent_fwspec.param[0] = hwirq + priv->parent_irq_base;
+		parent_fwspec.param[1] = fwspec->param[1];
+	} else
+		return -EINVAL;
 
-	return -ENXIO;
+	return irq_domain_alloc_irqs_parent(domain, virq, nr_irqs,
+			&parent_fwspec);
 }
 
+static void xgene_gpio_sb_domain_free(struct irq_domain *domain,
+		unsigned int virq,
+		unsigned int nr_irqs)
+{
+	struct irq_data *d;
+	unsigned int i;
+
+	for (i = 0; i < nr_irqs; i++) {
+		d = irq_domain_get_irq_data(domain, virq + i);
+		irq_domain_reset_irq_data(d);
+	}
+}
+
+static const struct irq_domain_ops xgene_gpio_sb_domain_ops = {
+	.translate      = xgene_gpio_sb_domain_translate,
+	.alloc          = xgene_gpio_sb_domain_alloc,
+	.free           = xgene_gpio_sb_domain_free,
+	.activate	= xgene_gpio_sb_domain_activate,
+	.deactivate	= xgene_gpio_sb_domain_deactivate,
+};
+
 static int xgene_gpio_sb_probe(struct platform_device *pdev)
 {
 	struct xgene_gpio_sb *priv;
-	u32 ret, i;
-	u32 default_lines[] = {0x08, 0x09, 0x0A, 0x0B, 0x0C, 0x0D};
+	u32 ret;
 	struct resource *res;
 	void __iomem *regs;
+	struct irq_domain *parent_domain = NULL;
+	struct fwnode_handle *fwnode;
+	u32 val32;
 
 	priv = devm_kzalloc(&pdev->dev, sizeof(*priv), GFP_KERNEL);
 	if (!priv)
@@ -90,6 +266,18 @@ static int xgene_gpio_sb_probe(struct platform_device *pdev)
 	if (IS_ERR(regs))
 		return PTR_ERR(regs);
 
+	priv->regs = regs;
+
+	ret = platform_get_irq(pdev, 0);
+	if (ret > 0) {
+		priv->parent_irq_base = irq_get_irq_data(ret)->hwirq;
+		parent_domain = irq_get_irq_data(ret)->domain;
+	}
+	if (!parent_domain) {
+		dev_err(&pdev->dev, "unable to obtain parent domain\n");
+		return -ENODEV;
+	}
+
 	ret = bgpio_init(&priv->gc, &pdev->dev, 4,
 			regs + MPA_GPIO_IN_ADDR,
 			regs + MPA_GPIO_OUT_ADDR, NULL,
@@ -97,30 +285,51 @@ static int xgene_gpio_sb_probe(struct platform_device *pdev)
         if (ret)
                 return ret;
 
-	priv->gc.to_irq = apm_gpio_sb_to_irq;
-	priv->gc.ngpio = XGENE_MAX_GPIO_DS;
+	priv->gc.to_irq = xgene_gpio_sb_to_irq;
 
-	priv->nirq = XGENE_MAX_GPIO_DS_IRQ;
+	/* Retrieve start irq pin, use default if property not found */
+	priv->irq_start = XGENE_DFLT_IRQ_START_PIN;
+	if (!device_property_read_u32(&pdev->dev,
+					XGENE_IRQ_START_PROPERTY, &val32))
+		priv->irq_start = val32;
 
-	priv->irq = devm_kzalloc(&pdev->dev, sizeof(u32) * XGENE_MAX_GPIO_DS,
-				   GFP_KERNEL);
-	if (!priv->irq)
-		return -ENOMEM;
+	/* Retrieve number irqs, use default if property not found */
+	priv->nirq = XGENE_DFLT_MAX_NIRQ;
+	if (!device_property_read_u32(&pdev->dev, XGENE_NIRQ_PROPERTY, &val32))
+		priv->nirq = val32;
 
-	for (i = 0; i < priv->nirq; i++) {
-		priv->irq[default_lines[i]] = platform_get_irq(pdev, i);
-		xgene_gpio_set_bit(&priv->gc, regs + MPA_GPIO_SEL_LO,
-                                   default_lines[i] * 2, 1);
-		xgene_gpio_set_bit(&priv->gc, regs + MPA_GPIO_INT_LVL, i, 1);
-	}
+	/* Retrieve number gpio, use default if property not found */
+	priv->gc.ngpio = XGENE_DFLT_MAX_NGPIO;
+	if (!device_property_read_u32(&pdev->dev, XGENE_NGPIO_PROPERTY, &val32))
+		priv->gc.ngpio = val32;
+
+	dev_info(&pdev->dev, "Support %d gpios, %d irqs start from pin %d\n",
+			priv->gc.ngpio, priv->nirq, priv->irq_start);
 
 	platform_set_drvdata(pdev, priv);
 
-	ret = gpiochip_add_data(&priv->gc, priv);
-	if (ret)
-		dev_err(&pdev->dev, "failed to register X-Gene GPIO Standby driver\n");
+	if (pdev->dev.of_node)
+		fwnode = of_node_to_fwnode(pdev->dev.of_node);
 	else
-		dev_info(&pdev->dev, "X-Gene GPIO Standby driver registered\n");
+		fwnode = pdev->dev.fwnode;
+
+	priv->irq_domain = irq_domain_create_hierarchy(parent_domain,
+					0, priv->nirq, fwnode,
+					&xgene_gpio_sb_domain_ops, priv);
+	if (!priv->irq_domain)
+		return -ENODEV;
+
+	priv->gc.irqdomain = priv->irq_domain;
+
+	ret = gpiochip_add_data(&priv->gc, priv);
+	if (ret) {
+		dev_err(&pdev->dev,
+			"failed to register X-Gene GPIO Standby driver\n");
+		irq_domain_remove(priv->irq_domain);
+		return ret;
+	}
+
+	dev_info(&pdev->dev, "X-Gene GPIO Standby driver registered\n");
 
 	if (priv->nirq > 0) {
 		/* Register interrupt handlers for gpio signaled acpi events */
@@ -138,6 +347,8 @@ static int xgene_gpio_sb_remove(struct platform_device *pdev)
 		acpi_gpiochip_free_interrupts(&priv->gc);
 	}
 
+	irq_domain_remove(priv->irq_domain);
+
 	gpiochip_remove(&priv->gc);
 	return 0;
 }
-- 
1.7.12.4

^ permalink raw reply related	[flat|nested] 24+ messages in thread

* [PATCH v5 2/3] Documentation: gpio: Update description for X-Gene standby GPIO controller DTS binding
  2016-01-29  4:28 ` Quan Nguyen
@ 2016-01-29  4:28   ` Quan Nguyen
  -1 siblings, 0 replies; 24+ messages in thread
From: Quan Nguyen @ 2016-01-29  4:28 UTC (permalink / raw)
  To: linus.walleij, linux-gpio, devicetree, linux-arm-kernel,
	Thomas Gleixner, Jason Cooper, Marc Zyngier
  Cc: Y Vo, Phong Vo, Loc Ho, Feng Kan, Duc Dang, patches, Quan Nguyen

Update description for X-Gene standby GPIO controller DTS binding to
support GPIO line configuration as input, output or external IRQ pin.

Signed-off-by: Y Vo <yvo@apm.com>
Signed-off-by: Quan Nguyen <qnguyen@apm.com>
---
 .../devicetree/bindings/gpio/gpio-xgene-sb.txt     | 47 ++++++++++++++++++----
 1 file changed, 40 insertions(+), 7 deletions(-)

diff --git a/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt b/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
index dae1300..7b8b4cb 100644
--- a/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
+++ b/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
@@ -1,10 +1,20 @@
 APM X-Gene Standby GPIO controller bindings
 
-This is a gpio controller in the standby domain.
-
-There are 20 GPIO pins from 0..21. There is no GPIO_DS14 or GPIO_DS15,
-only GPIO_DS8..GPIO_DS13 support interrupts. The IRQ mapping
-is currently 1-to-1 on interrupts 0x28 thru 0x2d.
+This is a gpio controller in the standby domain. It also supports interrupt in
+some particular pins which are sourced to its parent interrupt controller
+as diagram below:
+                           +-----------------+
+                           | X-Gene standby  |
+                           | GPIO controller +--------- GPIO_0
++------------+             |                 | ...
+| Parent IRQ |             |                 +--------- GPIO_8/EXT_INT_0
+| controller |  EXT_INT_0  |                 | ...
+| (GICv2)    +-------------+                 +--------- GPIO_[N+8]/EXT_INT_N
+|            |  ...        |                 |
+|            |  EXT_INT_N  |                 +--------- GPIO_[N+9]
+|            +-------------+                 | ...
+|            |             |                 +--------- GPIO_MAX
++------------+             +-----------------+
 
 Required properties:
 - compatible: "apm,xgene-gpio-sb" for the X-Gene Standby GPIO controller
@@ -15,10 +25,18 @@ Required properties:
 		0 = active high
 		1 = active low
 - gpio-controller: Marks the device node as a GPIO controller.
-- interrupts: Shall contain exactly 6 interrupts.
+- interrupts: The EXT_INT_0 parent interrupt resource must be listed first.
+- interrupt-parent: Phandle of the parent interrupt controller.
+- interrupt-cells: Should be two.
+       - first cell is 0-N coresponding for EXT_INT_0 to EXT_INT_N.
+       - second cell is used to specify flags.
+- interrupt-controller: Marks the device node as an interrupt controller.
+- apm,nr-gpios: Optional, specify number of gpios pin.
+- apm,nr-irqs: Optional, specify number of interrupt pins.
+- apm,irq-start: Optional, specify lowest gpio pin support interrupt.
 
 Example:
-	sbgpio: sbgpio@17001000 {
+	sbgpio: gpio@17001000{
 		compatible = "apm,xgene-gpio-sb";
 		reg = <0x0 0x17001000 0x0 0x400>;
 		#gpio-cells = <2>;
@@ -29,4 +47,19 @@ Example:
 				<0x0 0x2b 0x1>,
 				<0x0 0x2c 0x1>,
 				<0x0 0x2d 0x1>;
+		interrupt-parent = <&gic>;
+		#interrupt-cells = <2>;
+		interrupt-controller;
+		apm,nr-gpios = <22>;
+		apm,nr-irqs = <6>;
+		apm,irq-start = <8>;
+	};
+
+	testuser {
+		compatible = "example,testuser";
+		/* Use the GPIO_13/EXT_INT_5 line as an active high triggered
+		 * level interrupt
+		 */
+		interrupts = <5 4>;
+		interrupt-parent = <&sbgpio>;
 	};
-- 
1.7.12.4


^ permalink raw reply related	[flat|nested] 24+ messages in thread

* [PATCH v5 2/3] Documentation: gpio: Update description for X-Gene standby GPIO controller DTS binding
@ 2016-01-29  4:28   ` Quan Nguyen
  0 siblings, 0 replies; 24+ messages in thread
From: Quan Nguyen @ 2016-01-29  4:28 UTC (permalink / raw)
  To: linux-arm-kernel

Update description for X-Gene standby GPIO controller DTS binding to
support GPIO line configuration as input, output or external IRQ pin.

Signed-off-by: Y Vo <yvo@apm.com>
Signed-off-by: Quan Nguyen <qnguyen@apm.com>
---
 .../devicetree/bindings/gpio/gpio-xgene-sb.txt     | 47 ++++++++++++++++++----
 1 file changed, 40 insertions(+), 7 deletions(-)

diff --git a/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt b/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
index dae1300..7b8b4cb 100644
--- a/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
+++ b/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
@@ -1,10 +1,20 @@
 APM X-Gene Standby GPIO controller bindings
 
-This is a gpio controller in the standby domain.
-
-There are 20 GPIO pins from 0..21. There is no GPIO_DS14 or GPIO_DS15,
-only GPIO_DS8..GPIO_DS13 support interrupts. The IRQ mapping
-is currently 1-to-1 on interrupts 0x28 thru 0x2d.
+This is a gpio controller in the standby domain. It also supports interrupt in
+some particular pins which are sourced to its parent interrupt controller
+as diagram below:
+                           +-----------------+
+                           | X-Gene standby  |
+                           | GPIO controller +--------- GPIO_0
++------------+             |                 | ...
+| Parent IRQ |             |                 +--------- GPIO_8/EXT_INT_0
+| controller |  EXT_INT_0  |                 | ...
+| (GICv2)    +-------------+                 +--------- GPIO_[N+8]/EXT_INT_N
+|            |  ...        |                 |
+|            |  EXT_INT_N  |                 +--------- GPIO_[N+9]
+|            +-------------+                 | ...
+|            |             |                 +--------- GPIO_MAX
++------------+             +-----------------+
 
 Required properties:
 - compatible: "apm,xgene-gpio-sb" for the X-Gene Standby GPIO controller
@@ -15,10 +25,18 @@ Required properties:
 		0 = active high
 		1 = active low
 - gpio-controller: Marks the device node as a GPIO controller.
-- interrupts: Shall contain exactly 6 interrupts.
+- interrupts: The EXT_INT_0 parent interrupt resource must be listed first.
+- interrupt-parent: Phandle of the parent interrupt controller.
+- interrupt-cells: Should be two.
+       - first cell is 0-N coresponding for EXT_INT_0 to EXT_INT_N.
+       - second cell is used to specify flags.
+- interrupt-controller: Marks the device node as an interrupt controller.
+- apm,nr-gpios: Optional, specify number of gpios pin.
+- apm,nr-irqs: Optional, specify number of interrupt pins.
+- apm,irq-start: Optional, specify lowest gpio pin support interrupt.
 
 Example:
-	sbgpio: sbgpio at 17001000 {
+	sbgpio: gpio at 17001000{
 		compatible = "apm,xgene-gpio-sb";
 		reg = <0x0 0x17001000 0x0 0x400>;
 		#gpio-cells = <2>;
@@ -29,4 +47,19 @@ Example:
 				<0x0 0x2b 0x1>,
 				<0x0 0x2c 0x1>,
 				<0x0 0x2d 0x1>;
+		interrupt-parent = <&gic>;
+		#interrupt-cells = <2>;
+		interrupt-controller;
+		apm,nr-gpios = <22>;
+		apm,nr-irqs = <6>;
+		apm,irq-start = <8>;
+	};
+
+	testuser {
+		compatible = "example,testuser";
+		/* Use the GPIO_13/EXT_INT_5 line as an active high triggered
+		 * level interrupt
+		 */
+		interrupts = <5 4>;
+		interrupt-parent = <&sbgpio>;
 	};
-- 
1.7.12.4

^ permalink raw reply related	[flat|nested] 24+ messages in thread

* [PATCH v5 3/3] arm64: dts: Update X-Gene standby GPIO controller DTS entries
  2016-01-29  4:28 ` Quan Nguyen
@ 2016-01-29  4:28   ` Quan Nguyen
  -1 siblings, 0 replies; 24+ messages in thread
From: Quan Nguyen @ 2016-01-29  4:28 UTC (permalink / raw)
  To: linus.walleij, linux-gpio, devicetree, linux-arm-kernel,
	Thomas Gleixner, Jason Cooper, Marc Zyngier
  Cc: Y Vo, Phong Vo, Loc Ho, Feng Kan, Duc Dang, patches, Quan Nguyen

Update APM X-Gene standby GPIO controller DTS entries to enable it
as interrupt controller.

Signed-off-by: Y Vo <yvo@apm.com>
Signed-off-by: Quan Nguyen <qnguyen@apm.com>
---
 arch/arm64/boot/dts/apm/apm-storm.dtsi | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/arch/arm64/boot/dts/apm/apm-storm.dtsi b/arch/arm64/boot/dts/apm/apm-storm.dtsi
index fe30f76..0f733da 100644
--- a/arch/arm64/boot/dts/apm/apm-storm.dtsi
+++ b/arch/arm64/boot/dts/apm/apm-storm.dtsi
@@ -883,6 +883,9 @@
 					<0x0 0x2b 0x1>,
 					<0x0 0x2c 0x1>,
 					<0x0 0x2d 0x1>;
+			interrupt-parent = <&gic>;
+			#interrupt-cells = <2>;
+			interrupt-controller;
 		};
 
 		rtc: rtc@10510000 {
-- 
1.7.12.4


^ permalink raw reply related	[flat|nested] 24+ messages in thread

* [PATCH v5 3/3] arm64: dts: Update X-Gene standby GPIO controller DTS entries
@ 2016-01-29  4:28   ` Quan Nguyen
  0 siblings, 0 replies; 24+ messages in thread
From: Quan Nguyen @ 2016-01-29  4:28 UTC (permalink / raw)
  To: linux-arm-kernel

Update APM X-Gene standby GPIO controller DTS entries to enable it
as interrupt controller.

Signed-off-by: Y Vo <yvo@apm.com>
Signed-off-by: Quan Nguyen <qnguyen@apm.com>
---
 arch/arm64/boot/dts/apm/apm-storm.dtsi | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/arch/arm64/boot/dts/apm/apm-storm.dtsi b/arch/arm64/boot/dts/apm/apm-storm.dtsi
index fe30f76..0f733da 100644
--- a/arch/arm64/boot/dts/apm/apm-storm.dtsi
+++ b/arch/arm64/boot/dts/apm/apm-storm.dtsi
@@ -883,6 +883,9 @@
 					<0x0 0x2b 0x1>,
 					<0x0 0x2c 0x1>,
 					<0x0 0x2d 0x1>;
+			interrupt-parent = <&gic>;
+			#interrupt-cells = <2>;
+			interrupt-controller;
 		};
 
 		rtc: rtc at 10510000 {
-- 
1.7.12.4

^ permalink raw reply related	[flat|nested] 24+ messages in thread

* Re: [PATCH v5 2/3] Documentation: gpio: Update description for X-Gene standby GPIO controller DTS binding
  2016-01-29  4:28   ` Quan Nguyen
@ 2016-02-01 15:35       ` Rob Herring
  -1 siblings, 0 replies; 24+ messages in thread
From: Rob Herring @ 2016-02-01 15:35 UTC (permalink / raw)
  To: Quan Nguyen
  Cc: linus.walleij-QSEj5FYQhm4dnm+yROfE0A,
	linux-gpio-u79uwXL29TY76Z2rM5mHXA,
	devicetree-u79uwXL29TY76Z2rM5mHXA,
	linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r,
	Thomas Gleixner, Jason Cooper, Marc Zyngier, Y Vo, Phong Vo,
	Loc Ho, Feng Kan, Duc Dang, patches-qTEPVZfXA3Y

On Fri, Jan 29, 2016 at 11:28:54AM +0700, Quan Nguyen wrote:
> Update description for X-Gene standby GPIO controller DTS binding to
> support GPIO line configuration as input, output or external IRQ pin.
> 
> Signed-off-by: Y Vo <yvo-qTEPVZfXA3Y@public.gmane.org>
> Signed-off-by: Quan Nguyen <qnguyen-qTEPVZfXA3Y@public.gmane.org>
> ---
>  .../devicetree/bindings/gpio/gpio-xgene-sb.txt     | 47 ++++++++++++++++++----
>  1 file changed, 40 insertions(+), 7 deletions(-)
> 
> diff --git a/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt b/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
> index dae1300..7b8b4cb 100644
> --- a/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
> +++ b/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
> @@ -1,10 +1,20 @@
>  APM X-Gene Standby GPIO controller bindings
>  
> -This is a gpio controller in the standby domain.
> -
> -There are 20 GPIO pins from 0..21. There is no GPIO_DS14 or GPIO_DS15,
> -only GPIO_DS8..GPIO_DS13 support interrupts. The IRQ mapping
> -is currently 1-to-1 on interrupts 0x28 thru 0x2d.
> +This is a gpio controller in the standby domain. It also supports interrupt in
> +some particular pins which are sourced to its parent interrupt controller
> +as diagram below:
> +                           +-----------------+
> +                           | X-Gene standby  |
> +                           | GPIO controller +--------- GPIO_0
> ++------------+             |                 | ...
> +| Parent IRQ |             |                 +--------- GPIO_8/EXT_INT_0
> +| controller |  EXT_INT_0  |                 | ...
> +| (GICv2)    +-------------+                 +--------- GPIO_[N+8]/EXT_INT_N
> +|            |  ...        |                 |
> +|            |  EXT_INT_N  |                 +--------- GPIO_[N+9]
> +|            +-------------+                 | ...
> +|            |             |                 +--------- GPIO_MAX
> ++------------+             +-----------------+
>  
>  Required properties:
>  - compatible: "apm,xgene-gpio-sb" for the X-Gene Standby GPIO controller
> @@ -15,10 +25,18 @@ Required properties:
>  		0 = active high
>  		1 = active low
>  - gpio-controller: Marks the device node as a GPIO controller.
> -- interrupts: Shall contain exactly 6 interrupts.
> +- interrupts: The EXT_INT_0 parent interrupt resource must be listed first.
> +- interrupt-parent: Phandle of the parent interrupt controller.
> +- interrupt-cells: Should be two.
> +       - first cell is 0-N coresponding for EXT_INT_0 to EXT_INT_N.
> +       - second cell is used to specify flags.
> +- interrupt-controller: Marks the device node as an interrupt controller.
> +- apm,nr-gpios: Optional, specify number of gpios pin.
> +- apm,nr-irqs: Optional, specify number of interrupt pins.

When is this not 6?

> +- apm,irq-start: Optional, specify lowest gpio pin support interrupt.

What determines this value? What value is assumed if not present?

Rob
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

^ permalink raw reply	[flat|nested] 24+ messages in thread

* [PATCH v5 2/3] Documentation: gpio: Update description for X-Gene standby GPIO controller DTS binding
@ 2016-02-01 15:35       ` Rob Herring
  0 siblings, 0 replies; 24+ messages in thread
From: Rob Herring @ 2016-02-01 15:35 UTC (permalink / raw)
  To: linux-arm-kernel

On Fri, Jan 29, 2016 at 11:28:54AM +0700, Quan Nguyen wrote:
> Update description for X-Gene standby GPIO controller DTS binding to
> support GPIO line configuration as input, output or external IRQ pin.
> 
> Signed-off-by: Y Vo <yvo@apm.com>
> Signed-off-by: Quan Nguyen <qnguyen@apm.com>
> ---
>  .../devicetree/bindings/gpio/gpio-xgene-sb.txt     | 47 ++++++++++++++++++----
>  1 file changed, 40 insertions(+), 7 deletions(-)
> 
> diff --git a/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt b/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
> index dae1300..7b8b4cb 100644
> --- a/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
> +++ b/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
> @@ -1,10 +1,20 @@
>  APM X-Gene Standby GPIO controller bindings
>  
> -This is a gpio controller in the standby domain.
> -
> -There are 20 GPIO pins from 0..21. There is no GPIO_DS14 or GPIO_DS15,
> -only GPIO_DS8..GPIO_DS13 support interrupts. The IRQ mapping
> -is currently 1-to-1 on interrupts 0x28 thru 0x2d.
> +This is a gpio controller in the standby domain. It also supports interrupt in
> +some particular pins which are sourced to its parent interrupt controller
> +as diagram below:
> +                           +-----------------+
> +                           | X-Gene standby  |
> +                           | GPIO controller +--------- GPIO_0
> ++------------+             |                 | ...
> +| Parent IRQ |             |                 +--------- GPIO_8/EXT_INT_0
> +| controller |  EXT_INT_0  |                 | ...
> +| (GICv2)    +-------------+                 +--------- GPIO_[N+8]/EXT_INT_N
> +|            |  ...        |                 |
> +|            |  EXT_INT_N  |                 +--------- GPIO_[N+9]
> +|            +-------------+                 | ...
> +|            |             |                 +--------- GPIO_MAX
> ++------------+             +-----------------+
>  
>  Required properties:
>  - compatible: "apm,xgene-gpio-sb" for the X-Gene Standby GPIO controller
> @@ -15,10 +25,18 @@ Required properties:
>  		0 = active high
>  		1 = active low
>  - gpio-controller: Marks the device node as a GPIO controller.
> -- interrupts: Shall contain exactly 6 interrupts.
> +- interrupts: The EXT_INT_0 parent interrupt resource must be listed first.
> +- interrupt-parent: Phandle of the parent interrupt controller.
> +- interrupt-cells: Should be two.
> +       - first cell is 0-N coresponding for EXT_INT_0 to EXT_INT_N.
> +       - second cell is used to specify flags.
> +- interrupt-controller: Marks the device node as an interrupt controller.
> +- apm,nr-gpios: Optional, specify number of gpios pin.
> +- apm,nr-irqs: Optional, specify number of interrupt pins.

When is this not 6?

> +- apm,irq-start: Optional, specify lowest gpio pin support interrupt.

What determines this value? What value is assumed if not present?

Rob

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [PATCH v5 2/3] Documentation: gpio: Update description for X-Gene standby GPIO controller DTS binding
  2016-02-01 15:35       ` Rob Herring
@ 2016-02-02  1:46         ` Quan Nguyen
  -1 siblings, 0 replies; 24+ messages in thread
From: Quan Nguyen @ 2016-02-02  1:46 UTC (permalink / raw)
  To: Rob Herring
  Cc: Linus Walleij, linux-gpio-u79uwXL29TY76Z2rM5mHXA,
	devicetree-u79uwXL29TY76Z2rM5mHXA,
	linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r,
	Thomas Gleixner, Jason Cooper, Marc Zyngier, Y Vo, Phong Vo,
	Loc Ho, Feng Kan, Duc Dang, patches

On Mon, Feb 1, 2016 at 10:35 PM, Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote:
> On Fri, Jan 29, 2016 at 11:28:54AM +0700, Quan Nguyen wrote:
>> Update description for X-Gene standby GPIO controller DTS binding to
>> support GPIO line configuration as input, output or external IRQ pin.
>>
>> Signed-off-by: Y Vo <yvo-qTEPVZfXA3Y@public.gmane.org>
>> Signed-off-by: Quan Nguyen <qnguyen-qTEPVZfXA3Y@public.gmane.org>
>> ---
>>  .../devicetree/bindings/gpio/gpio-xgene-sb.txt     | 47 ++++++++++++++++++----
>>  1 file changed, 40 insertions(+), 7 deletions(-)
>>
>> diff --git a/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt b/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
>> index dae1300..7b8b4cb 100644
>> --- a/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
>> +++ b/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
>> @@ -1,10 +1,20 @@
>>  APM X-Gene Standby GPIO controller bindings
>>
>> -This is a gpio controller in the standby domain.
>> -
>> -There are 20 GPIO pins from 0..21. There is no GPIO_DS14 or GPIO_DS15,
>> -only GPIO_DS8..GPIO_DS13 support interrupts. The IRQ mapping
>> -is currently 1-to-1 on interrupts 0x28 thru 0x2d.
>> +This is a gpio controller in the standby domain. It also supports interrupt in
>> +some particular pins which are sourced to its parent interrupt controller
>> +as diagram below:
>> +                           +-----------------+
>> +                           | X-Gene standby  |
>> +                           | GPIO controller +--------- GPIO_0
>> ++------------+             |                 | ...
>> +| Parent IRQ |             |                 +--------- GPIO_8/EXT_INT_0
>> +| controller |  EXT_INT_0  |                 | ...
>> +| (GICv2)    +-------------+                 +--------- GPIO_[N+8]/EXT_INT_N
>> +|            |  ...        |                 |
>> +|            |  EXT_INT_N  |                 +--------- GPIO_[N+9]
>> +|            +-------------+                 | ...
>> +|            |             |                 +--------- GPIO_MAX
>> ++------------+             +-----------------+
>>
>>  Required properties:
>>  - compatible: "apm,xgene-gpio-sb" for the X-Gene Standby GPIO controller
>> @@ -15,10 +25,18 @@ Required properties:
>>               0 = active high
>>               1 = active low
>>  - gpio-controller: Marks the device node as a GPIO controller.
>> -- interrupts: Shall contain exactly 6 interrupts.
>> +- interrupts: The EXT_INT_0 parent interrupt resource must be listed first.
>> +- interrupt-parent: Phandle of the parent interrupt controller.
>> +- interrupt-cells: Should be two.
>> +       - first cell is 0-N coresponding for EXT_INT_0 to EXT_INT_N.
>> +       - second cell is used to specify flags.
>> +- interrupt-controller: Marks the device node as an interrupt controller.
>> +- apm,nr-gpios: Optional, specify number of gpios pin.
>> +- apm,nr-irqs: Optional, specify number of interrupt pins.
>
> When is this not 6?
>
Hi Rob,  by default, this should be 6, but I think this property can
help in cases:
+ Used only 5(or less) first pin as interrupt.
+ For similar device which has different interrupt pins (8 for example).
My idea is to make it a bit more generic by using this optional property.

>> +- apm,irq-start: Optional, specify lowest gpio pin support interrupt.
>
> What determines this value? What value is assumed if not present?
>
This value will tell interrupt is support from which gpio pin, as in
this case interrupt support from pin 8 (ie: only pin 8, 9,10 ...
supports interrupt) as default.
If not specify the default (ie: 8) will be used.

Thanks,
-- Quan Nguyen
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

^ permalink raw reply	[flat|nested] 24+ messages in thread

* [PATCH v5 2/3] Documentation: gpio: Update description for X-Gene standby GPIO controller DTS binding
@ 2016-02-02  1:46         ` Quan Nguyen
  0 siblings, 0 replies; 24+ messages in thread
From: Quan Nguyen @ 2016-02-02  1:46 UTC (permalink / raw)
  To: linux-arm-kernel

On Mon, Feb 1, 2016 at 10:35 PM, Rob Herring <robh@kernel.org> wrote:
> On Fri, Jan 29, 2016 at 11:28:54AM +0700, Quan Nguyen wrote:
>> Update description for X-Gene standby GPIO controller DTS binding to
>> support GPIO line configuration as input, output or external IRQ pin.
>>
>> Signed-off-by: Y Vo <yvo@apm.com>
>> Signed-off-by: Quan Nguyen <qnguyen@apm.com>
>> ---
>>  .../devicetree/bindings/gpio/gpio-xgene-sb.txt     | 47 ++++++++++++++++++----
>>  1 file changed, 40 insertions(+), 7 deletions(-)
>>
>> diff --git a/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt b/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
>> index dae1300..7b8b4cb 100644
>> --- a/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
>> +++ b/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
>> @@ -1,10 +1,20 @@
>>  APM X-Gene Standby GPIO controller bindings
>>
>> -This is a gpio controller in the standby domain.
>> -
>> -There are 20 GPIO pins from 0..21. There is no GPIO_DS14 or GPIO_DS15,
>> -only GPIO_DS8..GPIO_DS13 support interrupts. The IRQ mapping
>> -is currently 1-to-1 on interrupts 0x28 thru 0x2d.
>> +This is a gpio controller in the standby domain. It also supports interrupt in
>> +some particular pins which are sourced to its parent interrupt controller
>> +as diagram below:
>> +                           +-----------------+
>> +                           | X-Gene standby  |
>> +                           | GPIO controller +--------- GPIO_0
>> ++------------+             |                 | ...
>> +| Parent IRQ |             |                 +--------- GPIO_8/EXT_INT_0
>> +| controller |  EXT_INT_0  |                 | ...
>> +| (GICv2)    +-------------+                 +--------- GPIO_[N+8]/EXT_INT_N
>> +|            |  ...        |                 |
>> +|            |  EXT_INT_N  |                 +--------- GPIO_[N+9]
>> +|            +-------------+                 | ...
>> +|            |             |                 +--------- GPIO_MAX
>> ++------------+             +-----------------+
>>
>>  Required properties:
>>  - compatible: "apm,xgene-gpio-sb" for the X-Gene Standby GPIO controller
>> @@ -15,10 +25,18 @@ Required properties:
>>               0 = active high
>>               1 = active low
>>  - gpio-controller: Marks the device node as a GPIO controller.
>> -- interrupts: Shall contain exactly 6 interrupts.
>> +- interrupts: The EXT_INT_0 parent interrupt resource must be listed first.
>> +- interrupt-parent: Phandle of the parent interrupt controller.
>> +- interrupt-cells: Should be two.
>> +       - first cell is 0-N coresponding for EXT_INT_0 to EXT_INT_N.
>> +       - second cell is used to specify flags.
>> +- interrupt-controller: Marks the device node as an interrupt controller.
>> +- apm,nr-gpios: Optional, specify number of gpios pin.
>> +- apm,nr-irqs: Optional, specify number of interrupt pins.
>
> When is this not 6?
>
Hi Rob,  by default, this should be 6, but I think this property can
help in cases:
+ Used only 5(or less) first pin as interrupt.
+ For similar device which has different interrupt pins (8 for example).
My idea is to make it a bit more generic by using this optional property.

>> +- apm,irq-start: Optional, specify lowest gpio pin support interrupt.
>
> What determines this value? What value is assumed if not present?
>
This value will tell interrupt is support from which gpio pin, as in
this case interrupt support from pin 8 (ie: only pin 8, 9,10 ...
supports interrupt) as default.
If not specify the default (ie: 8) will be used.

Thanks,
-- Quan Nguyen

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [PATCH v5 1/3] gpio: xgene: Enable X-Gene standby GPIO as interrupt controller
  2016-01-29  4:28   ` Quan Nguyen
@ 2016-02-10 10:58       ` Linus Walleij
  -1 siblings, 0 replies; 24+ messages in thread
From: Linus Walleij @ 2016-02-10 10:58 UTC (permalink / raw)
  To: Quan Nguyen, Marc Zyngier
  Cc: linux-gpio-u79uwXL29TY76Z2rM5mHXA,
	devicetree-u79uwXL29TY76Z2rM5mHXA,
	linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r,
	Thomas Gleixner, Jason Cooper, Y Vo, Phong Vo, Loc Ho, Feng Kan,
	Duc Dang, patches

On Fri, Jan 29, 2016 at 5:28 AM, Quan Nguyen <qnguyen-qTEPVZfXA3Y@public.gmane.org> wrote:

> Enable X-Gene standby GPIO controller as interrupt controller to provide
> its own resources. This avoids ambiguity where GIC interrupt resource is
> use as X-Gene standby GPIO interrupt resource in user driver.
>
> Signed-off-by: Y Vo <yvo-qTEPVZfXA3Y@public.gmane.org>
> Signed-off-by: Quan Nguyen <qnguyen-qTEPVZfXA3Y@public.gmane.org>

I want Marc's ACK on this before I merge it.

But it is looking way more acceptable than earlier incarnations! :)

> +       /* Propagate IRQ type setting to parent */
> +       if (type & IRQ_TYPE_EDGE_BOTH)
> +               return irq_chip_set_type_parent(d, IRQ_TYPE_EDGE_RISING);
> +       else
> +               return irq_chip_set_type_parent(d, IRQ_TYPE_LEVEL_HIGH);

This makes it look much more solid, nice.

> +static void xgene_gpio_sb_irq_shutdown(struct irq_data *d)
> +{
> +       struct xgene_gpio_sb *priv = irq_data_get_irq_chip_data(d);
> +
> +       gpiochip_unlock_as_irq(&priv->gc, HWIRQ_TO_GPIO(priv, d->hwirq));

Oh really. Well.

> +       if (is_of_node(domain->parent->fwnode)) {
> +               parent_fwspec.fwnode = domain->parent->fwnode;
> +               parent_fwspec.param_count = 3;
> +               parent_fwspec.param[0] = 0;/* SPI */
> +               /* Skip SGIs and PPIs*/
> +               parent_fwspec.param[1] = hwirq + priv->parent_irq_base - 32;
> +               parent_fwspec.param[2] = fwspec->param[1];
> +       } else if (is_fwnode_irqchip(domain->parent->fwnode)) {
> +               parent_fwspec.fwnode = domain->parent->fwnode;
> +               parent_fwspec.param_count = 2;
> +               parent_fwspec.param[0] = hwirq + priv->parent_irq_base;
> +               parent_fwspec.param[1] = fwspec->param[1];
> +       } else
> +               return -EINVAL;

This kind of stuff hardcodes knowledge of the GIC into this controller,
like that the SGI+PPI occupy the first 32 hwIRQs. But maybe there is
no better way to do it?

Yours,
Linus Walleij
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

^ permalink raw reply	[flat|nested] 24+ messages in thread

* [PATCH v5 1/3] gpio: xgene: Enable X-Gene standby GPIO as interrupt controller
@ 2016-02-10 10:58       ` Linus Walleij
  0 siblings, 0 replies; 24+ messages in thread
From: Linus Walleij @ 2016-02-10 10:58 UTC (permalink / raw)
  To: linux-arm-kernel

On Fri, Jan 29, 2016 at 5:28 AM, Quan Nguyen <qnguyen@apm.com> wrote:

> Enable X-Gene standby GPIO controller as interrupt controller to provide
> its own resources. This avoids ambiguity where GIC interrupt resource is
> use as X-Gene standby GPIO interrupt resource in user driver.
>
> Signed-off-by: Y Vo <yvo@apm.com>
> Signed-off-by: Quan Nguyen <qnguyen@apm.com>

I want Marc's ACK on this before I merge it.

But it is looking way more acceptable than earlier incarnations! :)

> +       /* Propagate IRQ type setting to parent */
> +       if (type & IRQ_TYPE_EDGE_BOTH)
> +               return irq_chip_set_type_parent(d, IRQ_TYPE_EDGE_RISING);
> +       else
> +               return irq_chip_set_type_parent(d, IRQ_TYPE_LEVEL_HIGH);

This makes it look much more solid, nice.

> +static void xgene_gpio_sb_irq_shutdown(struct irq_data *d)
> +{
> +       struct xgene_gpio_sb *priv = irq_data_get_irq_chip_data(d);
> +
> +       gpiochip_unlock_as_irq(&priv->gc, HWIRQ_TO_GPIO(priv, d->hwirq));

Oh really. Well.

> +       if (is_of_node(domain->parent->fwnode)) {
> +               parent_fwspec.fwnode = domain->parent->fwnode;
> +               parent_fwspec.param_count = 3;
> +               parent_fwspec.param[0] = 0;/* SPI */
> +               /* Skip SGIs and PPIs*/
> +               parent_fwspec.param[1] = hwirq + priv->parent_irq_base - 32;
> +               parent_fwspec.param[2] = fwspec->param[1];
> +       } else if (is_fwnode_irqchip(domain->parent->fwnode)) {
> +               parent_fwspec.fwnode = domain->parent->fwnode;
> +               parent_fwspec.param_count = 2;
> +               parent_fwspec.param[0] = hwirq + priv->parent_irq_base;
> +               parent_fwspec.param[1] = fwspec->param[1];
> +       } else
> +               return -EINVAL;

This kind of stuff hardcodes knowledge of the GIC into this controller,
like that the SGI+PPI occupy the first 32 hwIRQs. But maybe there is
no better way to do it?

Yours,
Linus Walleij

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [PATCH v5 1/3] gpio: xgene: Enable X-Gene standby GPIO as interrupt controller
  2016-01-29  4:28   ` Quan Nguyen
@ 2016-02-10 15:09     ` Marc Zyngier
  -1 siblings, 0 replies; 24+ messages in thread
From: Marc Zyngier @ 2016-02-10 15:09 UTC (permalink / raw)
  To: Quan Nguyen, linus.walleij, linux-gpio, devicetree,
	linux-arm-kernel, Thomas Gleixner, Jason Cooper
  Cc: Y Vo, Phong Vo, Loc Ho, Feng Kan, Duc Dang, patches

Quan,

This is getting better. Some comments below.

On 29/01/16 04:28, Quan Nguyen wrote:
> Enable X-Gene standby GPIO controller as interrupt controller to provide
> its own resources. This avoids ambiguity where GIC interrupt resource is
> use as X-Gene standby GPIO interrupt resource in user driver.
> 
> Signed-off-by: Y Vo <yvo@apm.com>
> Signed-off-by: Quan Nguyen <qnguyen@apm.com>
> ---
>  drivers/gpio/gpio-xgene-sb.c | 275 ++++++++++++++++++++++++++++++++++++++-----
>  1 file changed, 243 insertions(+), 32 deletions(-)
> 
> diff --git a/drivers/gpio/gpio-xgene-sb.c b/drivers/gpio/gpio-xgene-sb.c
> index 282004d..026da90 100644
> --- a/drivers/gpio/gpio-xgene-sb.c
> +++ b/drivers/gpio/gpio-xgene-sb.c
> @@ -2,8 +2,9 @@
>   * AppliedMicro X-Gene SoC GPIO-Standby Driver
>   *
>   * Copyright (c) 2014, Applied Micro Circuits Corporation
> - * Author: 	Tin Huynh <tnhuynh@apm.com>.
> - * 		Y Vo <yvo@apm.com>.
> + * Author:	Tin Huynh <tnhuynh@apm.com>.
> + *		Y Vo <yvo@apm.com>.
> + *		Quan Nguyen <qnguyen@apm.com>.
>   *
>   * This program is free software; you can redistribute  it and/or modify it
>   * under  the terms of  the GNU General  Public License as published by the
> @@ -28,9 +29,14 @@
>  
>  #include "gpiolib.h"
>  
> -#define XGENE_MAX_GPIO_DS		22
> -#define XGENE_MAX_GPIO_DS_IRQ		6
> +/* Common property names */
> +#define XGENE_NIRQ_PROPERTY		"apm,nr-irqs"
> +#define XGENE_NGPIO_PROPERTY		"apm,nr-gpios"
> +#define XGENE_IRQ_START_PROPERTY	"apm,irq-start"
>  
> +#define XGENE_DFLT_MAX_NGPIO		22
> +#define XGENE_DFLT_MAX_NIRQ		6
> +#define XGENE_DFLT_IRQ_START_PIN	8
>  #define GPIO_MASK(x)			(1U << ((x) % 32))
>  
>  #define MPA_GPIO_INT_LVL		0x0290
> @@ -39,19 +45,32 @@
>  #define MPA_GPIO_IN_ADDR 		0x02a4
>  #define MPA_GPIO_SEL_LO 		0x0294
>  
> +#define GPIO_INT_LEVEL_H	0x000001
> +#define GPIO_INT_LEVEL_L	0x000000
> +
>  /**
>   * struct xgene_gpio_sb - GPIO-Standby private data structure.
>   * @gc:				memory-mapped GPIO controllers.
> - * @irq:			Mapping GPIO pins and interrupt number
> - * nirq:			Number of GPIO pins that supports interrupt
> + * @regs:			GPIO register base offset
> + * @irq_domain:			GPIO interrupt domain
> + * @irq_start:			GPIO pin that start support interrupt
> + * @nirq:			Number of GPIO pins that supports interrupt
> + * @parent_irq_base:		Start parent HWIRQ
>   */
>  struct xgene_gpio_sb {
>  	struct gpio_chip	gc;
> -	u32 *irq;
> -	u32 nirq;
> +	void __iomem		*regs;
> +	struct irq_domain	*irq_domain;
> +	u16			irq_start;
> +	u16			nirq;
> +	u16			parent_irq_base;
>  };
>  
> -static void xgene_gpio_set_bit(struct gpio_chip *gc, void __iomem *reg, u32 gpio, int val)
> +#define HWIRQ_TO_GPIO(priv, hwirq) ((hwirq) + (priv)->irq_start)
> +#define GPIO_TO_HWIRQ(priv, gpio) ((gpio) - (priv)->irq_start)
> +
> +static void xgene_gpio_set_bit(struct gpio_chip *gc,
> +				void __iomem *reg, u32 gpio, int val)
>  {
>  	u32 data;
>  
> @@ -63,23 +82,180 @@ static void xgene_gpio_set_bit(struct gpio_chip *gc, void __iomem *reg, u32 gpio
>  	gc->write_reg(reg, data);
>  }
>  
> -static int apm_gpio_sb_to_irq(struct gpio_chip *gc, u32 gpio)
> +static int xgene_gpio_sb_irq_set_type(struct irq_data *d, unsigned int type)
> +{
> +	struct xgene_gpio_sb *priv = irq_data_get_irq_chip_data(d);
> +	int gpio = HWIRQ_TO_GPIO(priv, d->hwirq);
> +	int lvl_type = GPIO_INT_LEVEL_H;
> +
> +	switch (type & IRQ_TYPE_SENSE_MASK) {
> +	case IRQ_TYPE_EDGE_RISING:
> +	case IRQ_TYPE_LEVEL_HIGH:
> +		lvl_type = GPIO_INT_LEVEL_H;
> +		break;
> +	case IRQ_TYPE_EDGE_FALLING:
> +	case IRQ_TYPE_LEVEL_LOW:
> +		lvl_type = GPIO_INT_LEVEL_L;
> +		break;
> +	default:
> +		break;
> +	}
> +
> +	xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_SEL_LO,
> +			gpio * 2, 1);
> +	xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_INT_LVL,
> +			d->hwirq, lvl_type);
> +
> +	/* Propagate IRQ type setting to parent */
> +	if (type & IRQ_TYPE_EDGE_BOTH)
> +		return irq_chip_set_type_parent(d, IRQ_TYPE_EDGE_RISING);
> +	else
> +		return irq_chip_set_type_parent(d, IRQ_TYPE_LEVEL_HIGH);
> +}
> +
> +static void xgene_gpio_sb_irq_shutdown(struct irq_data *d)
> +{
> +	struct xgene_gpio_sb *priv = irq_data_get_irq_chip_data(d);
> +
> +	gpiochip_unlock_as_irq(&priv->gc, HWIRQ_TO_GPIO(priv, d->hwirq));
> +}
> +
> +static struct irq_chip xgene_gpio_sb_irq_chip = {
> +	.name           = "sbgpio",
> +	.irq_ack        = irq_chip_ack_parent,

There is no point calling irq_chip_ack_parent, as the GIC doesn't
implement an irq_ack method at all.

> +	.irq_eoi	= irq_chip_eoi_parent,
> +	.irq_mask       = irq_chip_mask_parent,
> +	.irq_unmask     = irq_chip_unmask_parent,
> +	.irq_set_type   = xgene_gpio_sb_irq_set_type,
> +	.irq_shutdown   = xgene_gpio_sb_irq_shutdown,
> +};
> +
> +static int xgene_gpio_sb_to_irq(struct gpio_chip *gc, u32 gpio)
>  {
>  	struct xgene_gpio_sb *priv = gpiochip_get_data(gc);
> +	struct irq_fwspec fwspec;
> +
> +	if ((gpio < priv->irq_start) ||
> +			(gpio > HWIRQ_TO_GPIO(priv, priv->nirq)))
> +		return -ENXIO;
> +
> +	if (gc->parent->of_node)
> +		fwspec.fwnode = of_node_to_fwnode(gc->parent->of_node);
> +	else
> +		fwspec.fwnode = gc->parent->fwnode;
> +	fwspec.param_count = 2;
> +	fwspec.param[0] = GPIO_TO_HWIRQ(priv, gpio);
> +	fwspec.param[1] = IRQ_TYPE_NONE;
> +	return irq_create_fwspec_mapping(&fwspec);
> +}
> +
> +static void xgene_gpio_sb_domain_activate(struct irq_domain *d,
> +		struct irq_data *irq_data)
> +{
> +	struct xgene_gpio_sb *priv = d->host_data;
> +	u32 gpio = HWIRQ_TO_GPIO(priv, irq_data->hwirq);
> +
> +	if (gpiochip_lock_as_irq(&priv->gc, gpio)) {
> +		dev_err(priv->gc.parent,
> +		"Unable to configure XGene GPIO standby pin %d as IRQ\n",
> +				gpio);
> +		return;
> +	}
> +
> +	xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_SEL_LO,
> +			gpio * 2, 1);
> +}
> +
> +static void xgene_gpio_sb_domain_deactivate(struct irq_domain *d,
> +		struct irq_data *irq_data)
> +{
> +	struct xgene_gpio_sb *priv = d->host_data;
> +	u32 gpio = HWIRQ_TO_GPIO(priv, irq_data->hwirq);
> +
> +	gpiochip_unlock_as_irq(&priv->gc, gpio);

Is it right to do the unlock both in irq_shutdown and domain_deactivate?
This seems a bit odd to me to have such an inbalance. My hunch is that
you should either implement irq_startup, do the locking there and drop
the unlock drop deactivate, or kill irq_shutdown.

Linus, what do you think?

> +	xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_SEL_LO,
> +			gpio * 2, 0);
> +}
>  
> -	if (priv->irq[gpio])
> -		return priv->irq[gpio];
> +static int xgene_gpio_sb_domain_translate(struct irq_domain *d,
> +		struct irq_fwspec *fwspec,
> +		unsigned long *hwirq,
> +		unsigned int *type)
> +{
> +	struct xgene_gpio_sb *priv = d->host_data;
> +
> +	if ((fwspec->param_count != 2) ||
> +		(fwspec->param[0] >= priv->nirq))
> +		return -EINVAL;
> +	*hwirq = fwspec->param[0];
> +	*type = fwspec->param[1];
> +	return 0;
> +}
> +
> +static int xgene_gpio_sb_domain_alloc(struct irq_domain *domain,
> +					unsigned int virq,
> +					unsigned int nr_irqs, void *data)
> +{
> +	struct irq_fwspec *fwspec = data;
> +	struct irq_fwspec parent_fwspec;
> +	struct xgene_gpio_sb *priv = domain->host_data;
> +	irq_hw_number_t hwirq;
> +	unsigned int i;
> +
> +	hwirq = fwspec->param[0];
> +	for (i = 0; i < nr_irqs; i++)
> +		irq_domain_set_hwirq_and_chip(domain, virq + i, hwirq + i,
> +				&xgene_gpio_sb_irq_chip, priv);
> +
> +	if (is_of_node(domain->parent->fwnode)) {
> +		parent_fwspec.fwnode = domain->parent->fwnode;
> +		parent_fwspec.param_count = 3;
> +		parent_fwspec.param[0] = 0;/* SPI */
> +		/* Skip SGIs and PPIs*/
> +		parent_fwspec.param[1] = hwirq + priv->parent_irq_base - 32;
> +		parent_fwspec.param[2] = fwspec->param[1];
> +	} else if (is_fwnode_irqchip(domain->parent->fwnode)) {
> +		parent_fwspec.fwnode = domain->parent->fwnode;

nit: You can move this out of the if/else, and always assign the fwnode.

> +		parent_fwspec.param_count = 2;
> +		parent_fwspec.param[0] = hwirq + priv->parent_irq_base;
> +		parent_fwspec.param[1] = fwspec->param[1];
> +	} else
> +		return -EINVAL;
>  
> -	return -ENXIO;
> +	return irq_domain_alloc_irqs_parent(domain, virq, nr_irqs,
> +			&parent_fwspec);
>  }
>  
> +static void xgene_gpio_sb_domain_free(struct irq_domain *domain,
> +		unsigned int virq,
> +		unsigned int nr_irqs)
> +{
> +	struct irq_data *d;
> +	unsigned int i;
> +
> +	for (i = 0; i < nr_irqs; i++) {
> +		d = irq_domain_get_irq_data(domain, virq + i);
> +		irq_domain_reset_irq_data(d);
> +	}
> +}
> +
> +static const struct irq_domain_ops xgene_gpio_sb_domain_ops = {
> +	.translate      = xgene_gpio_sb_domain_translate,
> +	.alloc          = xgene_gpio_sb_domain_alloc,
> +	.free           = xgene_gpio_sb_domain_free,
> +	.activate	= xgene_gpio_sb_domain_activate,
> +	.deactivate	= xgene_gpio_sb_domain_deactivate,
> +};
> +
>  static int xgene_gpio_sb_probe(struct platform_device *pdev)
>  {
>  	struct xgene_gpio_sb *priv;
> -	u32 ret, i;
> -	u32 default_lines[] = {0x08, 0x09, 0x0A, 0x0B, 0x0C, 0x0D};
> +	u32 ret;
>  	struct resource *res;
>  	void __iomem *regs;
> +	struct irq_domain *parent_domain = NULL;
> +	struct fwnode_handle *fwnode;
> +	u32 val32;
>  
>  	priv = devm_kzalloc(&pdev->dev, sizeof(*priv), GFP_KERNEL);
>  	if (!priv)
> @@ -90,6 +266,18 @@ static int xgene_gpio_sb_probe(struct platform_device *pdev)
>  	if (IS_ERR(regs))
>  		return PTR_ERR(regs);
>  
> +	priv->regs = regs;
> +
> +	ret = platform_get_irq(pdev, 0);
> +	if (ret > 0) {
> +		priv->parent_irq_base = irq_get_irq_data(ret)->hwirq;
> +		parent_domain = irq_get_irq_data(ret)->domain;
> +	}
> +	if (!parent_domain) {
> +		dev_err(&pdev->dev, "unable to obtain parent domain\n");
> +		return -ENODEV;
> +	}
> +
>  	ret = bgpio_init(&priv->gc, &pdev->dev, 4,
>  			regs + MPA_GPIO_IN_ADDR,
>  			regs + MPA_GPIO_OUT_ADDR, NULL,
> @@ -97,30 +285,51 @@ static int xgene_gpio_sb_probe(struct platform_device *pdev)
>          if (ret)
>                  return ret;
>  
> -	priv->gc.to_irq = apm_gpio_sb_to_irq;
> -	priv->gc.ngpio = XGENE_MAX_GPIO_DS;
> +	priv->gc.to_irq = xgene_gpio_sb_to_irq;
>  
> -	priv->nirq = XGENE_MAX_GPIO_DS_IRQ;
> +	/* Retrieve start irq pin, use default if property not found */
> +	priv->irq_start = XGENE_DFLT_IRQ_START_PIN;
> +	if (!device_property_read_u32(&pdev->dev,
> +					XGENE_IRQ_START_PROPERTY, &val32))

Is that for ACPI as well?

> +		priv->irq_start = val32;
>  
> -	priv->irq = devm_kzalloc(&pdev->dev, sizeof(u32) * XGENE_MAX_GPIO_DS,
> -				   GFP_KERNEL);
> -	if (!priv->irq)
> -		return -ENOMEM;
> +	/* Retrieve number irqs, use default if property not found */
> +	priv->nirq = XGENE_DFLT_MAX_NIRQ;
> +	if (!device_property_read_u32(&pdev->dev, XGENE_NIRQ_PROPERTY, &val32))
> +		priv->nirq = val32;
>  
> -	for (i = 0; i < priv->nirq; i++) {
> -		priv->irq[default_lines[i]] = platform_get_irq(pdev, i);
> -		xgene_gpio_set_bit(&priv->gc, regs + MPA_GPIO_SEL_LO,
> -                                   default_lines[i] * 2, 1);
> -		xgene_gpio_set_bit(&priv->gc, regs + MPA_GPIO_INT_LVL, i, 1);
> -	}
> +	/* Retrieve number gpio, use default if property not found */
> +	priv->gc.ngpio = XGENE_DFLT_MAX_NGPIO;
> +	if (!device_property_read_u32(&pdev->dev, XGENE_NGPIO_PROPERTY, &val32))
> +		priv->gc.ngpio = val32;
> +
> +	dev_info(&pdev->dev, "Support %d gpios, %d irqs start from pin %d\n",
> +			priv->gc.ngpio, priv->nirq, priv->irq_start);
>  
>  	platform_set_drvdata(pdev, priv);
>  
> -	ret = gpiochip_add_data(&priv->gc, priv);
> -	if (ret)
> -		dev_err(&pdev->dev, "failed to register X-Gene GPIO Standby driver\n");
> +	if (pdev->dev.of_node)
> +		fwnode = of_node_to_fwnode(pdev->dev.of_node);
>  	else
> -		dev_info(&pdev->dev, "X-Gene GPIO Standby driver registered\n");
> +		fwnode = pdev->dev.fwnode;
> +
> +	priv->irq_domain = irq_domain_create_hierarchy(parent_domain,
> +					0, priv->nirq, fwnode,
> +					&xgene_gpio_sb_domain_ops, priv);
> +	if (!priv->irq_domain)
> +		return -ENODEV;
> +
> +	priv->gc.irqdomain = priv->irq_domain;
> +
> +	ret = gpiochip_add_data(&priv->gc, priv);
> +	if (ret) {
> +		dev_err(&pdev->dev,
> +			"failed to register X-Gene GPIO Standby driver\n");
> +		irq_domain_remove(priv->irq_domain);
> +		return ret;
> +	}
> +
> +	dev_info(&pdev->dev, "X-Gene GPIO Standby driver registered\n");
>  
>  	if (priv->nirq > 0) {
>  		/* Register interrupt handlers for gpio signaled acpi events */
> @@ -138,6 +347,8 @@ static int xgene_gpio_sb_remove(struct platform_device *pdev)
>  		acpi_gpiochip_free_interrupts(&priv->gc);
>  	}
>  
> +	irq_domain_remove(priv->irq_domain);
> +
>  	gpiochip_remove(&priv->gc);
>  	return 0;
>  }
> 

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...

^ permalink raw reply	[flat|nested] 24+ messages in thread

* [PATCH v5 1/3] gpio: xgene: Enable X-Gene standby GPIO as interrupt controller
@ 2016-02-10 15:09     ` Marc Zyngier
  0 siblings, 0 replies; 24+ messages in thread
From: Marc Zyngier @ 2016-02-10 15:09 UTC (permalink / raw)
  To: linux-arm-kernel

Quan,

This is getting better. Some comments below.

On 29/01/16 04:28, Quan Nguyen wrote:
> Enable X-Gene standby GPIO controller as interrupt controller to provide
> its own resources. This avoids ambiguity where GIC interrupt resource is
> use as X-Gene standby GPIO interrupt resource in user driver.
> 
> Signed-off-by: Y Vo <yvo@apm.com>
> Signed-off-by: Quan Nguyen <qnguyen@apm.com>
> ---
>  drivers/gpio/gpio-xgene-sb.c | 275 ++++++++++++++++++++++++++++++++++++++-----
>  1 file changed, 243 insertions(+), 32 deletions(-)
> 
> diff --git a/drivers/gpio/gpio-xgene-sb.c b/drivers/gpio/gpio-xgene-sb.c
> index 282004d..026da90 100644
> --- a/drivers/gpio/gpio-xgene-sb.c
> +++ b/drivers/gpio/gpio-xgene-sb.c
> @@ -2,8 +2,9 @@
>   * AppliedMicro X-Gene SoC GPIO-Standby Driver
>   *
>   * Copyright (c) 2014, Applied Micro Circuits Corporation
> - * Author: 	Tin Huynh <tnhuynh@apm.com>.
> - * 		Y Vo <yvo@apm.com>.
> + * Author:	Tin Huynh <tnhuynh@apm.com>.
> + *		Y Vo <yvo@apm.com>.
> + *		Quan Nguyen <qnguyen@apm.com>.
>   *
>   * This program is free software; you can redistribute  it and/or modify it
>   * under  the terms of  the GNU General  Public License as published by the
> @@ -28,9 +29,14 @@
>  
>  #include "gpiolib.h"
>  
> -#define XGENE_MAX_GPIO_DS		22
> -#define XGENE_MAX_GPIO_DS_IRQ		6
> +/* Common property names */
> +#define XGENE_NIRQ_PROPERTY		"apm,nr-irqs"
> +#define XGENE_NGPIO_PROPERTY		"apm,nr-gpios"
> +#define XGENE_IRQ_START_PROPERTY	"apm,irq-start"
>  
> +#define XGENE_DFLT_MAX_NGPIO		22
> +#define XGENE_DFLT_MAX_NIRQ		6
> +#define XGENE_DFLT_IRQ_START_PIN	8
>  #define GPIO_MASK(x)			(1U << ((x) % 32))
>  
>  #define MPA_GPIO_INT_LVL		0x0290
> @@ -39,19 +45,32 @@
>  #define MPA_GPIO_IN_ADDR 		0x02a4
>  #define MPA_GPIO_SEL_LO 		0x0294
>  
> +#define GPIO_INT_LEVEL_H	0x000001
> +#define GPIO_INT_LEVEL_L	0x000000
> +
>  /**
>   * struct xgene_gpio_sb - GPIO-Standby private data structure.
>   * @gc:				memory-mapped GPIO controllers.
> - * @irq:			Mapping GPIO pins and interrupt number
> - * nirq:			Number of GPIO pins that supports interrupt
> + * @regs:			GPIO register base offset
> + * @irq_domain:			GPIO interrupt domain
> + * @irq_start:			GPIO pin that start support interrupt
> + * @nirq:			Number of GPIO pins that supports interrupt
> + * @parent_irq_base:		Start parent HWIRQ
>   */
>  struct xgene_gpio_sb {
>  	struct gpio_chip	gc;
> -	u32 *irq;
> -	u32 nirq;
> +	void __iomem		*regs;
> +	struct irq_domain	*irq_domain;
> +	u16			irq_start;
> +	u16			nirq;
> +	u16			parent_irq_base;
>  };
>  
> -static void xgene_gpio_set_bit(struct gpio_chip *gc, void __iomem *reg, u32 gpio, int val)
> +#define HWIRQ_TO_GPIO(priv, hwirq) ((hwirq) + (priv)->irq_start)
> +#define GPIO_TO_HWIRQ(priv, gpio) ((gpio) - (priv)->irq_start)
> +
> +static void xgene_gpio_set_bit(struct gpio_chip *gc,
> +				void __iomem *reg, u32 gpio, int val)
>  {
>  	u32 data;
>  
> @@ -63,23 +82,180 @@ static void xgene_gpio_set_bit(struct gpio_chip *gc, void __iomem *reg, u32 gpio
>  	gc->write_reg(reg, data);
>  }
>  
> -static int apm_gpio_sb_to_irq(struct gpio_chip *gc, u32 gpio)
> +static int xgene_gpio_sb_irq_set_type(struct irq_data *d, unsigned int type)
> +{
> +	struct xgene_gpio_sb *priv = irq_data_get_irq_chip_data(d);
> +	int gpio = HWIRQ_TO_GPIO(priv, d->hwirq);
> +	int lvl_type = GPIO_INT_LEVEL_H;
> +
> +	switch (type & IRQ_TYPE_SENSE_MASK) {
> +	case IRQ_TYPE_EDGE_RISING:
> +	case IRQ_TYPE_LEVEL_HIGH:
> +		lvl_type = GPIO_INT_LEVEL_H;
> +		break;
> +	case IRQ_TYPE_EDGE_FALLING:
> +	case IRQ_TYPE_LEVEL_LOW:
> +		lvl_type = GPIO_INT_LEVEL_L;
> +		break;
> +	default:
> +		break;
> +	}
> +
> +	xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_SEL_LO,
> +			gpio * 2, 1);
> +	xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_INT_LVL,
> +			d->hwirq, lvl_type);
> +
> +	/* Propagate IRQ type setting to parent */
> +	if (type & IRQ_TYPE_EDGE_BOTH)
> +		return irq_chip_set_type_parent(d, IRQ_TYPE_EDGE_RISING);
> +	else
> +		return irq_chip_set_type_parent(d, IRQ_TYPE_LEVEL_HIGH);
> +}
> +
> +static void xgene_gpio_sb_irq_shutdown(struct irq_data *d)
> +{
> +	struct xgene_gpio_sb *priv = irq_data_get_irq_chip_data(d);
> +
> +	gpiochip_unlock_as_irq(&priv->gc, HWIRQ_TO_GPIO(priv, d->hwirq));
> +}
> +
> +static struct irq_chip xgene_gpio_sb_irq_chip = {
> +	.name           = "sbgpio",
> +	.irq_ack        = irq_chip_ack_parent,

There is no point calling irq_chip_ack_parent, as the GIC doesn't
implement an irq_ack method at all.

> +	.irq_eoi	= irq_chip_eoi_parent,
> +	.irq_mask       = irq_chip_mask_parent,
> +	.irq_unmask     = irq_chip_unmask_parent,
> +	.irq_set_type   = xgene_gpio_sb_irq_set_type,
> +	.irq_shutdown   = xgene_gpio_sb_irq_shutdown,
> +};
> +
> +static int xgene_gpio_sb_to_irq(struct gpio_chip *gc, u32 gpio)
>  {
>  	struct xgene_gpio_sb *priv = gpiochip_get_data(gc);
> +	struct irq_fwspec fwspec;
> +
> +	if ((gpio < priv->irq_start) ||
> +			(gpio > HWIRQ_TO_GPIO(priv, priv->nirq)))
> +		return -ENXIO;
> +
> +	if (gc->parent->of_node)
> +		fwspec.fwnode = of_node_to_fwnode(gc->parent->of_node);
> +	else
> +		fwspec.fwnode = gc->parent->fwnode;
> +	fwspec.param_count = 2;
> +	fwspec.param[0] = GPIO_TO_HWIRQ(priv, gpio);
> +	fwspec.param[1] = IRQ_TYPE_NONE;
> +	return irq_create_fwspec_mapping(&fwspec);
> +}
> +
> +static void xgene_gpio_sb_domain_activate(struct irq_domain *d,
> +		struct irq_data *irq_data)
> +{
> +	struct xgene_gpio_sb *priv = d->host_data;
> +	u32 gpio = HWIRQ_TO_GPIO(priv, irq_data->hwirq);
> +
> +	if (gpiochip_lock_as_irq(&priv->gc, gpio)) {
> +		dev_err(priv->gc.parent,
> +		"Unable to configure XGene GPIO standby pin %d as IRQ\n",
> +				gpio);
> +		return;
> +	}
> +
> +	xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_SEL_LO,
> +			gpio * 2, 1);
> +}
> +
> +static void xgene_gpio_sb_domain_deactivate(struct irq_domain *d,
> +		struct irq_data *irq_data)
> +{
> +	struct xgene_gpio_sb *priv = d->host_data;
> +	u32 gpio = HWIRQ_TO_GPIO(priv, irq_data->hwirq);
> +
> +	gpiochip_unlock_as_irq(&priv->gc, gpio);

Is it right to do the unlock both in irq_shutdown and domain_deactivate?
This seems a bit odd to me to have such an inbalance. My hunch is that
you should either implement irq_startup, do the locking there and drop
the unlock drop deactivate, or kill irq_shutdown.

Linus, what do you think?

> +	xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_SEL_LO,
> +			gpio * 2, 0);
> +}
>  
> -	if (priv->irq[gpio])
> -		return priv->irq[gpio];
> +static int xgene_gpio_sb_domain_translate(struct irq_domain *d,
> +		struct irq_fwspec *fwspec,
> +		unsigned long *hwirq,
> +		unsigned int *type)
> +{
> +	struct xgene_gpio_sb *priv = d->host_data;
> +
> +	if ((fwspec->param_count != 2) ||
> +		(fwspec->param[0] >= priv->nirq))
> +		return -EINVAL;
> +	*hwirq = fwspec->param[0];
> +	*type = fwspec->param[1];
> +	return 0;
> +}
> +
> +static int xgene_gpio_sb_domain_alloc(struct irq_domain *domain,
> +					unsigned int virq,
> +					unsigned int nr_irqs, void *data)
> +{
> +	struct irq_fwspec *fwspec = data;
> +	struct irq_fwspec parent_fwspec;
> +	struct xgene_gpio_sb *priv = domain->host_data;
> +	irq_hw_number_t hwirq;
> +	unsigned int i;
> +
> +	hwirq = fwspec->param[0];
> +	for (i = 0; i < nr_irqs; i++)
> +		irq_domain_set_hwirq_and_chip(domain, virq + i, hwirq + i,
> +				&xgene_gpio_sb_irq_chip, priv);
> +
> +	if (is_of_node(domain->parent->fwnode)) {
> +		parent_fwspec.fwnode = domain->parent->fwnode;
> +		parent_fwspec.param_count = 3;
> +		parent_fwspec.param[0] = 0;/* SPI */
> +		/* Skip SGIs and PPIs*/
> +		parent_fwspec.param[1] = hwirq + priv->parent_irq_base - 32;
> +		parent_fwspec.param[2] = fwspec->param[1];
> +	} else if (is_fwnode_irqchip(domain->parent->fwnode)) {
> +		parent_fwspec.fwnode = domain->parent->fwnode;

nit: You can move this out of the if/else, and always assign the fwnode.

> +		parent_fwspec.param_count = 2;
> +		parent_fwspec.param[0] = hwirq + priv->parent_irq_base;
> +		parent_fwspec.param[1] = fwspec->param[1];
> +	} else
> +		return -EINVAL;
>  
> -	return -ENXIO;
> +	return irq_domain_alloc_irqs_parent(domain, virq, nr_irqs,
> +			&parent_fwspec);
>  }
>  
> +static void xgene_gpio_sb_domain_free(struct irq_domain *domain,
> +		unsigned int virq,
> +		unsigned int nr_irqs)
> +{
> +	struct irq_data *d;
> +	unsigned int i;
> +
> +	for (i = 0; i < nr_irqs; i++) {
> +		d = irq_domain_get_irq_data(domain, virq + i);
> +		irq_domain_reset_irq_data(d);
> +	}
> +}
> +
> +static const struct irq_domain_ops xgene_gpio_sb_domain_ops = {
> +	.translate      = xgene_gpio_sb_domain_translate,
> +	.alloc          = xgene_gpio_sb_domain_alloc,
> +	.free           = xgene_gpio_sb_domain_free,
> +	.activate	= xgene_gpio_sb_domain_activate,
> +	.deactivate	= xgene_gpio_sb_domain_deactivate,
> +};
> +
>  static int xgene_gpio_sb_probe(struct platform_device *pdev)
>  {
>  	struct xgene_gpio_sb *priv;
> -	u32 ret, i;
> -	u32 default_lines[] = {0x08, 0x09, 0x0A, 0x0B, 0x0C, 0x0D};
> +	u32 ret;
>  	struct resource *res;
>  	void __iomem *regs;
> +	struct irq_domain *parent_domain = NULL;
> +	struct fwnode_handle *fwnode;
> +	u32 val32;
>  
>  	priv = devm_kzalloc(&pdev->dev, sizeof(*priv), GFP_KERNEL);
>  	if (!priv)
> @@ -90,6 +266,18 @@ static int xgene_gpio_sb_probe(struct platform_device *pdev)
>  	if (IS_ERR(regs))
>  		return PTR_ERR(regs);
>  
> +	priv->regs = regs;
> +
> +	ret = platform_get_irq(pdev, 0);
> +	if (ret > 0) {
> +		priv->parent_irq_base = irq_get_irq_data(ret)->hwirq;
> +		parent_domain = irq_get_irq_data(ret)->domain;
> +	}
> +	if (!parent_domain) {
> +		dev_err(&pdev->dev, "unable to obtain parent domain\n");
> +		return -ENODEV;
> +	}
> +
>  	ret = bgpio_init(&priv->gc, &pdev->dev, 4,
>  			regs + MPA_GPIO_IN_ADDR,
>  			regs + MPA_GPIO_OUT_ADDR, NULL,
> @@ -97,30 +285,51 @@ static int xgene_gpio_sb_probe(struct platform_device *pdev)
>          if (ret)
>                  return ret;
>  
> -	priv->gc.to_irq = apm_gpio_sb_to_irq;
> -	priv->gc.ngpio = XGENE_MAX_GPIO_DS;
> +	priv->gc.to_irq = xgene_gpio_sb_to_irq;
>  
> -	priv->nirq = XGENE_MAX_GPIO_DS_IRQ;
> +	/* Retrieve start irq pin, use default if property not found */
> +	priv->irq_start = XGENE_DFLT_IRQ_START_PIN;
> +	if (!device_property_read_u32(&pdev->dev,
> +					XGENE_IRQ_START_PROPERTY, &val32))

Is that for ACPI as well?

> +		priv->irq_start = val32;
>  
> -	priv->irq = devm_kzalloc(&pdev->dev, sizeof(u32) * XGENE_MAX_GPIO_DS,
> -				   GFP_KERNEL);
> -	if (!priv->irq)
> -		return -ENOMEM;
> +	/* Retrieve number irqs, use default if property not found */
> +	priv->nirq = XGENE_DFLT_MAX_NIRQ;
> +	if (!device_property_read_u32(&pdev->dev, XGENE_NIRQ_PROPERTY, &val32))
> +		priv->nirq = val32;
>  
> -	for (i = 0; i < priv->nirq; i++) {
> -		priv->irq[default_lines[i]] = platform_get_irq(pdev, i);
> -		xgene_gpio_set_bit(&priv->gc, regs + MPA_GPIO_SEL_LO,
> -                                   default_lines[i] * 2, 1);
> -		xgene_gpio_set_bit(&priv->gc, regs + MPA_GPIO_INT_LVL, i, 1);
> -	}
> +	/* Retrieve number gpio, use default if property not found */
> +	priv->gc.ngpio = XGENE_DFLT_MAX_NGPIO;
> +	if (!device_property_read_u32(&pdev->dev, XGENE_NGPIO_PROPERTY, &val32))
> +		priv->gc.ngpio = val32;
> +
> +	dev_info(&pdev->dev, "Support %d gpios, %d irqs start from pin %d\n",
> +			priv->gc.ngpio, priv->nirq, priv->irq_start);
>  
>  	platform_set_drvdata(pdev, priv);
>  
> -	ret = gpiochip_add_data(&priv->gc, priv);
> -	if (ret)
> -		dev_err(&pdev->dev, "failed to register X-Gene GPIO Standby driver\n");
> +	if (pdev->dev.of_node)
> +		fwnode = of_node_to_fwnode(pdev->dev.of_node);
>  	else
> -		dev_info(&pdev->dev, "X-Gene GPIO Standby driver registered\n");
> +		fwnode = pdev->dev.fwnode;
> +
> +	priv->irq_domain = irq_domain_create_hierarchy(parent_domain,
> +					0, priv->nirq, fwnode,
> +					&xgene_gpio_sb_domain_ops, priv);
> +	if (!priv->irq_domain)
> +		return -ENODEV;
> +
> +	priv->gc.irqdomain = priv->irq_domain;
> +
> +	ret = gpiochip_add_data(&priv->gc, priv);
> +	if (ret) {
> +		dev_err(&pdev->dev,
> +			"failed to register X-Gene GPIO Standby driver\n");
> +		irq_domain_remove(priv->irq_domain);
> +		return ret;
> +	}
> +
> +	dev_info(&pdev->dev, "X-Gene GPIO Standby driver registered\n");
>  
>  	if (priv->nirq > 0) {
>  		/* Register interrupt handlers for gpio signaled acpi events */
> @@ -138,6 +347,8 @@ static int xgene_gpio_sb_remove(struct platform_device *pdev)
>  		acpi_gpiochip_free_interrupts(&priv->gc);
>  	}
>  
> +	irq_domain_remove(priv->irq_domain);
> +
>  	gpiochip_remove(&priv->gc);
>  	return 0;
>  }
> 

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [PATCH v5 1/3] gpio: xgene: Enable X-Gene standby GPIO as interrupt controller
  2016-02-10 15:09     ` Marc Zyngier
@ 2016-02-11 16:35         ` Quan Nguyen
  -1 siblings, 0 replies; 24+ messages in thread
From: Quan Nguyen @ 2016-02-11 16:35 UTC (permalink / raw)
  To: Marc Zyngier
  Cc: Linus Walleij, linux-gpio-u79uwXL29TY76Z2rM5mHXA,
	devicetree-u79uwXL29TY76Z2rM5mHXA,
	linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r,
	Thomas Gleixner, Jason Cooper, Y Vo, Phong Vo, Loc Ho, Feng Kan,
	Duc Dang, patches

On Wed, Feb 10, 2016 at 10:09 PM, Marc Zyngier <marc.zyngier-5wv7dgnIgG8@public.gmane.org> wrote:
> Quan,
>
> This is getting better. Some comments below.
>
> On 29/01/16 04:28, Quan Nguyen wrote:
>> Enable X-Gene standby GPIO controller as interrupt controller to provide
>> its own resources. This avoids ambiguity where GIC interrupt resource is
>> use as X-Gene standby GPIO interrupt resource in user driver.
>>
>> Signed-off-by: Y Vo <yvo-qTEPVZfXA3Y@public.gmane.org>
>> Signed-off-by: Quan Nguyen <qnguyen-qTEPVZfXA3Y@public.gmane.org>
>> ---
>>  drivers/gpio/gpio-xgene-sb.c | 275 ++++++++++++++++++++++++++++++++++++++-----
>>  1 file changed, 243 insertions(+), 32 deletions(-)
>>
>> diff --git a/drivers/gpio/gpio-xgene-sb.c b/drivers/gpio/gpio-xgene-sb.c
>> index 282004d..026da90 100644
>> --- a/drivers/gpio/gpio-xgene-sb.c
>> +++ b/drivers/gpio/gpio-xgene-sb.c
>> @@ -2,8 +2,9 @@
>>   * AppliedMicro X-Gene SoC GPIO-Standby Driver
>>   *
>>   * Copyright (c) 2014, Applied Micro Circuits Corporation
>> - * Author:   Tin Huynh <tnhuynh-qTEPVZfXA3Y@public.gmane.org>.
>> - *           Y Vo <yvo-qTEPVZfXA3Y@public.gmane.org>.
>> + * Author:   Tin Huynh <tnhuynh-qTEPVZfXA3Y@public.gmane.org>.
>> + *           Y Vo <yvo-qTEPVZfXA3Y@public.gmane.org>.
>> + *           Quan Nguyen <qnguyen-qTEPVZfXA3Y@public.gmane.org>.
>>   *
>>   * This program is free software; you can redistribute  it and/or modify it
>>   * under  the terms of  the GNU General  Public License as published by the
>> @@ -28,9 +29,14 @@
>>
>>  #include "gpiolib.h"
>>
>> -#define XGENE_MAX_GPIO_DS            22
>> -#define XGENE_MAX_GPIO_DS_IRQ                6
>> +/* Common property names */
>> +#define XGENE_NIRQ_PROPERTY          "apm,nr-irqs"
>> +#define XGENE_NGPIO_PROPERTY         "apm,nr-gpios"
>> +#define XGENE_IRQ_START_PROPERTY     "apm,irq-start"
>>
>> +#define XGENE_DFLT_MAX_NGPIO         22
>> +#define XGENE_DFLT_MAX_NIRQ          6
>> +#define XGENE_DFLT_IRQ_START_PIN     8
>>  #define GPIO_MASK(x)                 (1U << ((x) % 32))
>>
>>  #define MPA_GPIO_INT_LVL             0x0290
>> @@ -39,19 +45,32 @@
>>  #define MPA_GPIO_IN_ADDR             0x02a4
>>  #define MPA_GPIO_SEL_LO              0x0294
>>
>> +#define GPIO_INT_LEVEL_H     0x000001
>> +#define GPIO_INT_LEVEL_L     0x000000
>> +
>>  /**
>>   * struct xgene_gpio_sb - GPIO-Standby private data structure.
>>   * @gc:                              memory-mapped GPIO controllers.
>> - * @irq:                     Mapping GPIO pins and interrupt number
>> - * nirq:                     Number of GPIO pins that supports interrupt
>> + * @regs:                    GPIO register base offset
>> + * @irq_domain:                      GPIO interrupt domain
>> + * @irq_start:                       GPIO pin that start support interrupt
>> + * @nirq:                    Number of GPIO pins that supports interrupt
>> + * @parent_irq_base:         Start parent HWIRQ
>>   */
>>  struct xgene_gpio_sb {
>>       struct gpio_chip        gc;
>> -     u32 *irq;
>> -     u32 nirq;
>> +     void __iomem            *regs;
>> +     struct irq_domain       *irq_domain;
>> +     u16                     irq_start;
>> +     u16                     nirq;
>> +     u16                     parent_irq_base;
>>  };
>>
>> -static void xgene_gpio_set_bit(struct gpio_chip *gc, void __iomem *reg, u32 gpio, int val)
>> +#define HWIRQ_TO_GPIO(priv, hwirq) ((hwirq) + (priv)->irq_start)
>> +#define GPIO_TO_HWIRQ(priv, gpio) ((gpio) - (priv)->irq_start)
>> +
>> +static void xgene_gpio_set_bit(struct gpio_chip *gc,
>> +                             void __iomem *reg, u32 gpio, int val)
>>  {
>>       u32 data;
>>
>> @@ -63,23 +82,180 @@ static void xgene_gpio_set_bit(struct gpio_chip *gc, void __iomem *reg, u32 gpio
>>       gc->write_reg(reg, data);
>>  }
>>
>> -static int apm_gpio_sb_to_irq(struct gpio_chip *gc, u32 gpio)
>> +static int xgene_gpio_sb_irq_set_type(struct irq_data *d, unsigned int type)
>> +{
>> +     struct xgene_gpio_sb *priv = irq_data_get_irq_chip_data(d);
>> +     int gpio = HWIRQ_TO_GPIO(priv, d->hwirq);
>> +     int lvl_type = GPIO_INT_LEVEL_H;
>> +
>> +     switch (type & IRQ_TYPE_SENSE_MASK) {
>> +     case IRQ_TYPE_EDGE_RISING:
>> +     case IRQ_TYPE_LEVEL_HIGH:
>> +             lvl_type = GPIO_INT_LEVEL_H;
>> +             break;
>> +     case IRQ_TYPE_EDGE_FALLING:
>> +     case IRQ_TYPE_LEVEL_LOW:
>> +             lvl_type = GPIO_INT_LEVEL_L;
>> +             break;
>> +     default:
>> +             break;
>> +     }
>> +
>> +     xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_SEL_LO,
>> +                     gpio * 2, 1);
>> +     xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_INT_LVL,
>> +                     d->hwirq, lvl_type);
>> +
>> +     /* Propagate IRQ type setting to parent */
>> +     if (type & IRQ_TYPE_EDGE_BOTH)
>> +             return irq_chip_set_type_parent(d, IRQ_TYPE_EDGE_RISING);
>> +     else
>> +             return irq_chip_set_type_parent(d, IRQ_TYPE_LEVEL_HIGH);
>> +}
>> +
>> +static void xgene_gpio_sb_irq_shutdown(struct irq_data *d)
>> +{
>> +     struct xgene_gpio_sb *priv = irq_data_get_irq_chip_data(d);
>> +
>> +     gpiochip_unlock_as_irq(&priv->gc, HWIRQ_TO_GPIO(priv, d->hwirq));
>> +}
>> +
>> +static struct irq_chip xgene_gpio_sb_irq_chip = {
>> +     .name           = "sbgpio",
>> +     .irq_ack        = irq_chip_ack_parent,
>
> There is no point calling irq_chip_ack_parent, as the GIC doesn't
> implement an irq_ack method at all.
>

Thanks Marc, let me remove it.

>> +     .irq_eoi        = irq_chip_eoi_parent,
>> +     .irq_mask       = irq_chip_mask_parent,
>> +     .irq_unmask     = irq_chip_unmask_parent,
>> +     .irq_set_type   = xgene_gpio_sb_irq_set_type,
>> +     .irq_shutdown   = xgene_gpio_sb_irq_shutdown,
>> +};
>> +
>> +static int xgene_gpio_sb_to_irq(struct gpio_chip *gc, u32 gpio)
>>  {
>>       struct xgene_gpio_sb *priv = gpiochip_get_data(gc);
>> +     struct irq_fwspec fwspec;
>> +
>> +     if ((gpio < priv->irq_start) ||
>> +                     (gpio > HWIRQ_TO_GPIO(priv, priv->nirq)))
>> +             return -ENXIO;
>> +
>> +     if (gc->parent->of_node)
>> +             fwspec.fwnode = of_node_to_fwnode(gc->parent->of_node);
>> +     else
>> +             fwspec.fwnode = gc->parent->fwnode;
>> +     fwspec.param_count = 2;
>> +     fwspec.param[0] = GPIO_TO_HWIRQ(priv, gpio);
>> +     fwspec.param[1] = IRQ_TYPE_NONE;
>> +     return irq_create_fwspec_mapping(&fwspec);
>> +}
>> +
>> +static void xgene_gpio_sb_domain_activate(struct irq_domain *d,
>> +             struct irq_data *irq_data)
>> +{
>> +     struct xgene_gpio_sb *priv = d->host_data;
>> +     u32 gpio = HWIRQ_TO_GPIO(priv, irq_data->hwirq);
>> +
>> +     if (gpiochip_lock_as_irq(&priv->gc, gpio)) {
>> +             dev_err(priv->gc.parent,
>> +             "Unable to configure XGene GPIO standby pin %d as IRQ\n",
>> +                             gpio);
>> +             return;
>> +     }
>> +
>> +     xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_SEL_LO,
>> +                     gpio * 2, 1);
>> +}
>> +
>> +static void xgene_gpio_sb_domain_deactivate(struct irq_domain *d,
>> +             struct irq_data *irq_data)
>> +{
>> +     struct xgene_gpio_sb *priv = d->host_data;
>> +     u32 gpio = HWIRQ_TO_GPIO(priv, irq_data->hwirq);
>> +
>> +     gpiochip_unlock_as_irq(&priv->gc, gpio);
>
> Is it right to do the unlock both in irq_shutdown and domain_deactivate?
> This seems a bit odd to me to have such an inbalance. My hunch is that
> you should either implement irq_startup, do the locking there and drop
> the unlock drop deactivate, or kill irq_shutdown.
>
> Linus, what do you think?
>

I refer to kill irq_shutdown.

>> +     xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_SEL_LO,
>> +                     gpio * 2, 0);
>> +}
>>
>> -     if (priv->irq[gpio])
>> -             return priv->irq[gpio];
>> +static int xgene_gpio_sb_domain_translate(struct irq_domain *d,
>> +             struct irq_fwspec *fwspec,
>> +             unsigned long *hwirq,
>> +             unsigned int *type)
>> +{
>> +     struct xgene_gpio_sb *priv = d->host_data;
>> +
>> +     if ((fwspec->param_count != 2) ||
>> +             (fwspec->param[0] >= priv->nirq))
>> +             return -EINVAL;
>> +     *hwirq = fwspec->param[0];
>> +     *type = fwspec->param[1];
>> +     return 0;
>> +}
>> +
>> +static int xgene_gpio_sb_domain_alloc(struct irq_domain *domain,
>> +                                     unsigned int virq,
>> +                                     unsigned int nr_irqs, void *data)
>> +{
>> +     struct irq_fwspec *fwspec = data;
>> +     struct irq_fwspec parent_fwspec;
>> +     struct xgene_gpio_sb *priv = domain->host_data;
>> +     irq_hw_number_t hwirq;
>> +     unsigned int i;
>> +
>> +     hwirq = fwspec->param[0];
>> +     for (i = 0; i < nr_irqs; i++)
>> +             irq_domain_set_hwirq_and_chip(domain, virq + i, hwirq + i,
>> +                             &xgene_gpio_sb_irq_chip, priv);
>> +
>> +     if (is_of_node(domain->parent->fwnode)) {
>> +             parent_fwspec.fwnode = domain->parent->fwnode;
>> +             parent_fwspec.param_count = 3;
>> +             parent_fwspec.param[0] = 0;/* SPI */
>> +             /* Skip SGIs and PPIs*/
>> +             parent_fwspec.param[1] = hwirq + priv->parent_irq_base - 32;
>> +             parent_fwspec.param[2] = fwspec->param[1];
>> +     } else if (is_fwnode_irqchip(domain->parent->fwnode)) {
>> +             parent_fwspec.fwnode = domain->parent->fwnode;
>
> nit: You can move this out of the if/else, and always assign the fwnode.

Thanks Marc, let me change it.

>
>> +             parent_fwspec.param_count = 2;
>> +             parent_fwspec.param[0] = hwirq + priv->parent_irq_base;
>> +             parent_fwspec.param[1] = fwspec->param[1];
>> +     } else
>> +             return -EINVAL;
>>
>> -     return -ENXIO;
>> +     return irq_domain_alloc_irqs_parent(domain, virq, nr_irqs,
>> +                     &parent_fwspec);
>>  }
>>
>> +static void xgene_gpio_sb_domain_free(struct irq_domain *domain,
>> +             unsigned int virq,
>> +             unsigned int nr_irqs)
>> +{
>> +     struct irq_data *d;
>> +     unsigned int i;
>> +
>> +     for (i = 0; i < nr_irqs; i++) {
>> +             d = irq_domain_get_irq_data(domain, virq + i);
>> +             irq_domain_reset_irq_data(d);
>> +     }
>> +}
>> +
>> +static const struct irq_domain_ops xgene_gpio_sb_domain_ops = {
>> +     .translate      = xgene_gpio_sb_domain_translate,
>> +     .alloc          = xgene_gpio_sb_domain_alloc,
>> +     .free           = xgene_gpio_sb_domain_free,
>> +     .activate       = xgene_gpio_sb_domain_activate,
>> +     .deactivate     = xgene_gpio_sb_domain_deactivate,
>> +};
>> +
>>  static int xgene_gpio_sb_probe(struct platform_device *pdev)
>>  {
>>       struct xgene_gpio_sb *priv;
>> -     u32 ret, i;
>> -     u32 default_lines[] = {0x08, 0x09, 0x0A, 0x0B, 0x0C, 0x0D};
>> +     u32 ret;
>>       struct resource *res;
>>       void __iomem *regs;
>> +     struct irq_domain *parent_domain = NULL;
>> +     struct fwnode_handle *fwnode;
>> +     u32 val32;
>>
>>       priv = devm_kzalloc(&pdev->dev, sizeof(*priv), GFP_KERNEL);
>>       if (!priv)
>> @@ -90,6 +266,18 @@ static int xgene_gpio_sb_probe(struct platform_device *pdev)
>>       if (IS_ERR(regs))
>>               return PTR_ERR(regs);
>>
>> +     priv->regs = regs;
>> +
>> +     ret = platform_get_irq(pdev, 0);
>> +     if (ret > 0) {
>> +             priv->parent_irq_base = irq_get_irq_data(ret)->hwirq;
>> +             parent_domain = irq_get_irq_data(ret)->domain;
>> +     }
>> +     if (!parent_domain) {
>> +             dev_err(&pdev->dev, "unable to obtain parent domain\n");
>> +             return -ENODEV;
>> +     }
>> +
>>       ret = bgpio_init(&priv->gc, &pdev->dev, 4,
>>                       regs + MPA_GPIO_IN_ADDR,
>>                       regs + MPA_GPIO_OUT_ADDR, NULL,
>> @@ -97,30 +285,51 @@ static int xgene_gpio_sb_probe(struct platform_device *pdev)
>>          if (ret)
>>                  return ret;
>>
>> -     priv->gc.to_irq = apm_gpio_sb_to_irq;
>> -     priv->gc.ngpio = XGENE_MAX_GPIO_DS;
>> +     priv->gc.to_irq = xgene_gpio_sb_to_irq;
>>
>> -     priv->nirq = XGENE_MAX_GPIO_DS_IRQ;
>> +     /* Retrieve start irq pin, use default if property not found */
>> +     priv->irq_start = XGENE_DFLT_IRQ_START_PIN;
>> +     if (!device_property_read_u32(&pdev->dev,
>> +                                     XGENE_IRQ_START_PROPERTY, &val32))
>
> Is that for ACPI as well?
>

Yes, Marc.

>> +             priv->irq_start = val32;
>>
>> -     priv->irq = devm_kzalloc(&pdev->dev, sizeof(u32) * XGENE_MAX_GPIO_DS,
>> -                                GFP_KERNEL);
>> -     if (!priv->irq)
>> -             return -ENOMEM;
>> +     /* Retrieve number irqs, use default if property not found */
>> +     priv->nirq = XGENE_DFLT_MAX_NIRQ;
>> +     if (!device_property_read_u32(&pdev->dev, XGENE_NIRQ_PROPERTY, &val32))
>> +             priv->nirq = val32;
>>
>> -     for (i = 0; i < priv->nirq; i++) {
>> -             priv->irq[default_lines[i]] = platform_get_irq(pdev, i);
>> -             xgene_gpio_set_bit(&priv->gc, regs + MPA_GPIO_SEL_LO,
>> -                                   default_lines[i] * 2, 1);
>> -             xgene_gpio_set_bit(&priv->gc, regs + MPA_GPIO_INT_LVL, i, 1);
>> -     }
>> +     /* Retrieve number gpio, use default if property not found */
>> +     priv->gc.ngpio = XGENE_DFLT_MAX_NGPIO;
>> +     if (!device_property_read_u32(&pdev->dev, XGENE_NGPIO_PROPERTY, &val32))
>> +             priv->gc.ngpio = val32;
>> +
>> +     dev_info(&pdev->dev, "Support %d gpios, %d irqs start from pin %d\n",
>> +                     priv->gc.ngpio, priv->nirq, priv->irq_start);
>>
>>       platform_set_drvdata(pdev, priv);
>>
>> -     ret = gpiochip_add_data(&priv->gc, priv);
>> -     if (ret)
>> -             dev_err(&pdev->dev, "failed to register X-Gene GPIO Standby driver\n");
>> +     if (pdev->dev.of_node)
>> +             fwnode = of_node_to_fwnode(pdev->dev.of_node);
>>       else
>> -             dev_info(&pdev->dev, "X-Gene GPIO Standby driver registered\n");
>> +             fwnode = pdev->dev.fwnode;
>> +
>> +     priv->irq_domain = irq_domain_create_hierarchy(parent_domain,
>> +                                     0, priv->nirq, fwnode,
>> +                                     &xgene_gpio_sb_domain_ops, priv);
>> +     if (!priv->irq_domain)
>> +             return -ENODEV;
>> +
>> +     priv->gc.irqdomain = priv->irq_domain;
>> +
>> +     ret = gpiochip_add_data(&priv->gc, priv);
>> +     if (ret) {
>> +             dev_err(&pdev->dev,
>> +                     "failed to register X-Gene GPIO Standby driver\n");
>> +             irq_domain_remove(priv->irq_domain);
>> +             return ret;
>> +     }
>> +
>> +     dev_info(&pdev->dev, "X-Gene GPIO Standby driver registered\n");
>>
>>       if (priv->nirq > 0) {
>>               /* Register interrupt handlers for gpio signaled acpi events */
>> @@ -138,6 +347,8 @@ static int xgene_gpio_sb_remove(struct platform_device *pdev)
>>               acpi_gpiochip_free_interrupts(&priv->gc);
>>       }
>>
>> +     irq_domain_remove(priv->irq_domain);
>> +
>>       gpiochip_remove(&priv->gc);
>>       return 0;
>>  }
>>
>
> Thanks,
>
>         M.
> --
> Jazz is not dead. It just smells funny...
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

^ permalink raw reply	[flat|nested] 24+ messages in thread

* [PATCH v5 1/3] gpio: xgene: Enable X-Gene standby GPIO as interrupt controller
@ 2016-02-11 16:35         ` Quan Nguyen
  0 siblings, 0 replies; 24+ messages in thread
From: Quan Nguyen @ 2016-02-11 16:35 UTC (permalink / raw)
  To: linux-arm-kernel

On Wed, Feb 10, 2016 at 10:09 PM, Marc Zyngier <marc.zyngier@arm.com> wrote:
> Quan,
>
> This is getting better. Some comments below.
>
> On 29/01/16 04:28, Quan Nguyen wrote:
>> Enable X-Gene standby GPIO controller as interrupt controller to provide
>> its own resources. This avoids ambiguity where GIC interrupt resource is
>> use as X-Gene standby GPIO interrupt resource in user driver.
>>
>> Signed-off-by: Y Vo <yvo@apm.com>
>> Signed-off-by: Quan Nguyen <qnguyen@apm.com>
>> ---
>>  drivers/gpio/gpio-xgene-sb.c | 275 ++++++++++++++++++++++++++++++++++++++-----
>>  1 file changed, 243 insertions(+), 32 deletions(-)
>>
>> diff --git a/drivers/gpio/gpio-xgene-sb.c b/drivers/gpio/gpio-xgene-sb.c
>> index 282004d..026da90 100644
>> --- a/drivers/gpio/gpio-xgene-sb.c
>> +++ b/drivers/gpio/gpio-xgene-sb.c
>> @@ -2,8 +2,9 @@
>>   * AppliedMicro X-Gene SoC GPIO-Standby Driver
>>   *
>>   * Copyright (c) 2014, Applied Micro Circuits Corporation
>> - * Author:   Tin Huynh <tnhuynh@apm.com>.
>> - *           Y Vo <yvo@apm.com>.
>> + * Author:   Tin Huynh <tnhuynh@apm.com>.
>> + *           Y Vo <yvo@apm.com>.
>> + *           Quan Nguyen <qnguyen@apm.com>.
>>   *
>>   * This program is free software; you can redistribute  it and/or modify it
>>   * under  the terms of  the GNU General  Public License as published by the
>> @@ -28,9 +29,14 @@
>>
>>  #include "gpiolib.h"
>>
>> -#define XGENE_MAX_GPIO_DS            22
>> -#define XGENE_MAX_GPIO_DS_IRQ                6
>> +/* Common property names */
>> +#define XGENE_NIRQ_PROPERTY          "apm,nr-irqs"
>> +#define XGENE_NGPIO_PROPERTY         "apm,nr-gpios"
>> +#define XGENE_IRQ_START_PROPERTY     "apm,irq-start"
>>
>> +#define XGENE_DFLT_MAX_NGPIO         22
>> +#define XGENE_DFLT_MAX_NIRQ          6
>> +#define XGENE_DFLT_IRQ_START_PIN     8
>>  #define GPIO_MASK(x)                 (1U << ((x) % 32))
>>
>>  #define MPA_GPIO_INT_LVL             0x0290
>> @@ -39,19 +45,32 @@
>>  #define MPA_GPIO_IN_ADDR             0x02a4
>>  #define MPA_GPIO_SEL_LO              0x0294
>>
>> +#define GPIO_INT_LEVEL_H     0x000001
>> +#define GPIO_INT_LEVEL_L     0x000000
>> +
>>  /**
>>   * struct xgene_gpio_sb - GPIO-Standby private data structure.
>>   * @gc:                              memory-mapped GPIO controllers.
>> - * @irq:                     Mapping GPIO pins and interrupt number
>> - * nirq:                     Number of GPIO pins that supports interrupt
>> + * @regs:                    GPIO register base offset
>> + * @irq_domain:                      GPIO interrupt domain
>> + * @irq_start:                       GPIO pin that start support interrupt
>> + * @nirq:                    Number of GPIO pins that supports interrupt
>> + * @parent_irq_base:         Start parent HWIRQ
>>   */
>>  struct xgene_gpio_sb {
>>       struct gpio_chip        gc;
>> -     u32 *irq;
>> -     u32 nirq;
>> +     void __iomem            *regs;
>> +     struct irq_domain       *irq_domain;
>> +     u16                     irq_start;
>> +     u16                     nirq;
>> +     u16                     parent_irq_base;
>>  };
>>
>> -static void xgene_gpio_set_bit(struct gpio_chip *gc, void __iomem *reg, u32 gpio, int val)
>> +#define HWIRQ_TO_GPIO(priv, hwirq) ((hwirq) + (priv)->irq_start)
>> +#define GPIO_TO_HWIRQ(priv, gpio) ((gpio) - (priv)->irq_start)
>> +
>> +static void xgene_gpio_set_bit(struct gpio_chip *gc,
>> +                             void __iomem *reg, u32 gpio, int val)
>>  {
>>       u32 data;
>>
>> @@ -63,23 +82,180 @@ static void xgene_gpio_set_bit(struct gpio_chip *gc, void __iomem *reg, u32 gpio
>>       gc->write_reg(reg, data);
>>  }
>>
>> -static int apm_gpio_sb_to_irq(struct gpio_chip *gc, u32 gpio)
>> +static int xgene_gpio_sb_irq_set_type(struct irq_data *d, unsigned int type)
>> +{
>> +     struct xgene_gpio_sb *priv = irq_data_get_irq_chip_data(d);
>> +     int gpio = HWIRQ_TO_GPIO(priv, d->hwirq);
>> +     int lvl_type = GPIO_INT_LEVEL_H;
>> +
>> +     switch (type & IRQ_TYPE_SENSE_MASK) {
>> +     case IRQ_TYPE_EDGE_RISING:
>> +     case IRQ_TYPE_LEVEL_HIGH:
>> +             lvl_type = GPIO_INT_LEVEL_H;
>> +             break;
>> +     case IRQ_TYPE_EDGE_FALLING:
>> +     case IRQ_TYPE_LEVEL_LOW:
>> +             lvl_type = GPIO_INT_LEVEL_L;
>> +             break;
>> +     default:
>> +             break;
>> +     }
>> +
>> +     xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_SEL_LO,
>> +                     gpio * 2, 1);
>> +     xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_INT_LVL,
>> +                     d->hwirq, lvl_type);
>> +
>> +     /* Propagate IRQ type setting to parent */
>> +     if (type & IRQ_TYPE_EDGE_BOTH)
>> +             return irq_chip_set_type_parent(d, IRQ_TYPE_EDGE_RISING);
>> +     else
>> +             return irq_chip_set_type_parent(d, IRQ_TYPE_LEVEL_HIGH);
>> +}
>> +
>> +static void xgene_gpio_sb_irq_shutdown(struct irq_data *d)
>> +{
>> +     struct xgene_gpio_sb *priv = irq_data_get_irq_chip_data(d);
>> +
>> +     gpiochip_unlock_as_irq(&priv->gc, HWIRQ_TO_GPIO(priv, d->hwirq));
>> +}
>> +
>> +static struct irq_chip xgene_gpio_sb_irq_chip = {
>> +     .name           = "sbgpio",
>> +     .irq_ack        = irq_chip_ack_parent,
>
> There is no point calling irq_chip_ack_parent, as the GIC doesn't
> implement an irq_ack method at all.
>

Thanks Marc, let me remove it.

>> +     .irq_eoi        = irq_chip_eoi_parent,
>> +     .irq_mask       = irq_chip_mask_parent,
>> +     .irq_unmask     = irq_chip_unmask_parent,
>> +     .irq_set_type   = xgene_gpio_sb_irq_set_type,
>> +     .irq_shutdown   = xgene_gpio_sb_irq_shutdown,
>> +};
>> +
>> +static int xgene_gpio_sb_to_irq(struct gpio_chip *gc, u32 gpio)
>>  {
>>       struct xgene_gpio_sb *priv = gpiochip_get_data(gc);
>> +     struct irq_fwspec fwspec;
>> +
>> +     if ((gpio < priv->irq_start) ||
>> +                     (gpio > HWIRQ_TO_GPIO(priv, priv->nirq)))
>> +             return -ENXIO;
>> +
>> +     if (gc->parent->of_node)
>> +             fwspec.fwnode = of_node_to_fwnode(gc->parent->of_node);
>> +     else
>> +             fwspec.fwnode = gc->parent->fwnode;
>> +     fwspec.param_count = 2;
>> +     fwspec.param[0] = GPIO_TO_HWIRQ(priv, gpio);
>> +     fwspec.param[1] = IRQ_TYPE_NONE;
>> +     return irq_create_fwspec_mapping(&fwspec);
>> +}
>> +
>> +static void xgene_gpio_sb_domain_activate(struct irq_domain *d,
>> +             struct irq_data *irq_data)
>> +{
>> +     struct xgene_gpio_sb *priv = d->host_data;
>> +     u32 gpio = HWIRQ_TO_GPIO(priv, irq_data->hwirq);
>> +
>> +     if (gpiochip_lock_as_irq(&priv->gc, gpio)) {
>> +             dev_err(priv->gc.parent,
>> +             "Unable to configure XGene GPIO standby pin %d as IRQ\n",
>> +                             gpio);
>> +             return;
>> +     }
>> +
>> +     xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_SEL_LO,
>> +                     gpio * 2, 1);
>> +}
>> +
>> +static void xgene_gpio_sb_domain_deactivate(struct irq_domain *d,
>> +             struct irq_data *irq_data)
>> +{
>> +     struct xgene_gpio_sb *priv = d->host_data;
>> +     u32 gpio = HWIRQ_TO_GPIO(priv, irq_data->hwirq);
>> +
>> +     gpiochip_unlock_as_irq(&priv->gc, gpio);
>
> Is it right to do the unlock both in irq_shutdown and domain_deactivate?
> This seems a bit odd to me to have such an inbalance. My hunch is that
> you should either implement irq_startup, do the locking there and drop
> the unlock drop deactivate, or kill irq_shutdown.
>
> Linus, what do you think?
>

I refer to kill irq_shutdown.

>> +     xgene_gpio_set_bit(&priv->gc, priv->regs + MPA_GPIO_SEL_LO,
>> +                     gpio * 2, 0);
>> +}
>>
>> -     if (priv->irq[gpio])
>> -             return priv->irq[gpio];
>> +static int xgene_gpio_sb_domain_translate(struct irq_domain *d,
>> +             struct irq_fwspec *fwspec,
>> +             unsigned long *hwirq,
>> +             unsigned int *type)
>> +{
>> +     struct xgene_gpio_sb *priv = d->host_data;
>> +
>> +     if ((fwspec->param_count != 2) ||
>> +             (fwspec->param[0] >= priv->nirq))
>> +             return -EINVAL;
>> +     *hwirq = fwspec->param[0];
>> +     *type = fwspec->param[1];
>> +     return 0;
>> +}
>> +
>> +static int xgene_gpio_sb_domain_alloc(struct irq_domain *domain,
>> +                                     unsigned int virq,
>> +                                     unsigned int nr_irqs, void *data)
>> +{
>> +     struct irq_fwspec *fwspec = data;
>> +     struct irq_fwspec parent_fwspec;
>> +     struct xgene_gpio_sb *priv = domain->host_data;
>> +     irq_hw_number_t hwirq;
>> +     unsigned int i;
>> +
>> +     hwirq = fwspec->param[0];
>> +     for (i = 0; i < nr_irqs; i++)
>> +             irq_domain_set_hwirq_and_chip(domain, virq + i, hwirq + i,
>> +                             &xgene_gpio_sb_irq_chip, priv);
>> +
>> +     if (is_of_node(domain->parent->fwnode)) {
>> +             parent_fwspec.fwnode = domain->parent->fwnode;
>> +             parent_fwspec.param_count = 3;
>> +             parent_fwspec.param[0] = 0;/* SPI */
>> +             /* Skip SGIs and PPIs*/
>> +             parent_fwspec.param[1] = hwirq + priv->parent_irq_base - 32;
>> +             parent_fwspec.param[2] = fwspec->param[1];
>> +     } else if (is_fwnode_irqchip(domain->parent->fwnode)) {
>> +             parent_fwspec.fwnode = domain->parent->fwnode;
>
> nit: You can move this out of the if/else, and always assign the fwnode.

Thanks Marc, let me change it.

>
>> +             parent_fwspec.param_count = 2;
>> +             parent_fwspec.param[0] = hwirq + priv->parent_irq_base;
>> +             parent_fwspec.param[1] = fwspec->param[1];
>> +     } else
>> +             return -EINVAL;
>>
>> -     return -ENXIO;
>> +     return irq_domain_alloc_irqs_parent(domain, virq, nr_irqs,
>> +                     &parent_fwspec);
>>  }
>>
>> +static void xgene_gpio_sb_domain_free(struct irq_domain *domain,
>> +             unsigned int virq,
>> +             unsigned int nr_irqs)
>> +{
>> +     struct irq_data *d;
>> +     unsigned int i;
>> +
>> +     for (i = 0; i < nr_irqs; i++) {
>> +             d = irq_domain_get_irq_data(domain, virq + i);
>> +             irq_domain_reset_irq_data(d);
>> +     }
>> +}
>> +
>> +static const struct irq_domain_ops xgene_gpio_sb_domain_ops = {
>> +     .translate      = xgene_gpio_sb_domain_translate,
>> +     .alloc          = xgene_gpio_sb_domain_alloc,
>> +     .free           = xgene_gpio_sb_domain_free,
>> +     .activate       = xgene_gpio_sb_domain_activate,
>> +     .deactivate     = xgene_gpio_sb_domain_deactivate,
>> +};
>> +
>>  static int xgene_gpio_sb_probe(struct platform_device *pdev)
>>  {
>>       struct xgene_gpio_sb *priv;
>> -     u32 ret, i;
>> -     u32 default_lines[] = {0x08, 0x09, 0x0A, 0x0B, 0x0C, 0x0D};
>> +     u32 ret;
>>       struct resource *res;
>>       void __iomem *regs;
>> +     struct irq_domain *parent_domain = NULL;
>> +     struct fwnode_handle *fwnode;
>> +     u32 val32;
>>
>>       priv = devm_kzalloc(&pdev->dev, sizeof(*priv), GFP_KERNEL);
>>       if (!priv)
>> @@ -90,6 +266,18 @@ static int xgene_gpio_sb_probe(struct platform_device *pdev)
>>       if (IS_ERR(regs))
>>               return PTR_ERR(regs);
>>
>> +     priv->regs = regs;
>> +
>> +     ret = platform_get_irq(pdev, 0);
>> +     if (ret > 0) {
>> +             priv->parent_irq_base = irq_get_irq_data(ret)->hwirq;
>> +             parent_domain = irq_get_irq_data(ret)->domain;
>> +     }
>> +     if (!parent_domain) {
>> +             dev_err(&pdev->dev, "unable to obtain parent domain\n");
>> +             return -ENODEV;
>> +     }
>> +
>>       ret = bgpio_init(&priv->gc, &pdev->dev, 4,
>>                       regs + MPA_GPIO_IN_ADDR,
>>                       regs + MPA_GPIO_OUT_ADDR, NULL,
>> @@ -97,30 +285,51 @@ static int xgene_gpio_sb_probe(struct platform_device *pdev)
>>          if (ret)
>>                  return ret;
>>
>> -     priv->gc.to_irq = apm_gpio_sb_to_irq;
>> -     priv->gc.ngpio = XGENE_MAX_GPIO_DS;
>> +     priv->gc.to_irq = xgene_gpio_sb_to_irq;
>>
>> -     priv->nirq = XGENE_MAX_GPIO_DS_IRQ;
>> +     /* Retrieve start irq pin, use default if property not found */
>> +     priv->irq_start = XGENE_DFLT_IRQ_START_PIN;
>> +     if (!device_property_read_u32(&pdev->dev,
>> +                                     XGENE_IRQ_START_PROPERTY, &val32))
>
> Is that for ACPI as well?
>

Yes, Marc.

>> +             priv->irq_start = val32;
>>
>> -     priv->irq = devm_kzalloc(&pdev->dev, sizeof(u32) * XGENE_MAX_GPIO_DS,
>> -                                GFP_KERNEL);
>> -     if (!priv->irq)
>> -             return -ENOMEM;
>> +     /* Retrieve number irqs, use default if property not found */
>> +     priv->nirq = XGENE_DFLT_MAX_NIRQ;
>> +     if (!device_property_read_u32(&pdev->dev, XGENE_NIRQ_PROPERTY, &val32))
>> +             priv->nirq = val32;
>>
>> -     for (i = 0; i < priv->nirq; i++) {
>> -             priv->irq[default_lines[i]] = platform_get_irq(pdev, i);
>> -             xgene_gpio_set_bit(&priv->gc, regs + MPA_GPIO_SEL_LO,
>> -                                   default_lines[i] * 2, 1);
>> -             xgene_gpio_set_bit(&priv->gc, regs + MPA_GPIO_INT_LVL, i, 1);
>> -     }
>> +     /* Retrieve number gpio, use default if property not found */
>> +     priv->gc.ngpio = XGENE_DFLT_MAX_NGPIO;
>> +     if (!device_property_read_u32(&pdev->dev, XGENE_NGPIO_PROPERTY, &val32))
>> +             priv->gc.ngpio = val32;
>> +
>> +     dev_info(&pdev->dev, "Support %d gpios, %d irqs start from pin %d\n",
>> +                     priv->gc.ngpio, priv->nirq, priv->irq_start);
>>
>>       platform_set_drvdata(pdev, priv);
>>
>> -     ret = gpiochip_add_data(&priv->gc, priv);
>> -     if (ret)
>> -             dev_err(&pdev->dev, "failed to register X-Gene GPIO Standby driver\n");
>> +     if (pdev->dev.of_node)
>> +             fwnode = of_node_to_fwnode(pdev->dev.of_node);
>>       else
>> -             dev_info(&pdev->dev, "X-Gene GPIO Standby driver registered\n");
>> +             fwnode = pdev->dev.fwnode;
>> +
>> +     priv->irq_domain = irq_domain_create_hierarchy(parent_domain,
>> +                                     0, priv->nirq, fwnode,
>> +                                     &xgene_gpio_sb_domain_ops, priv);
>> +     if (!priv->irq_domain)
>> +             return -ENODEV;
>> +
>> +     priv->gc.irqdomain = priv->irq_domain;
>> +
>> +     ret = gpiochip_add_data(&priv->gc, priv);
>> +     if (ret) {
>> +             dev_err(&pdev->dev,
>> +                     "failed to register X-Gene GPIO Standby driver\n");
>> +             irq_domain_remove(priv->irq_domain);
>> +             return ret;
>> +     }
>> +
>> +     dev_info(&pdev->dev, "X-Gene GPIO Standby driver registered\n");
>>
>>       if (priv->nirq > 0) {
>>               /* Register interrupt handlers for gpio signaled acpi events */
>> @@ -138,6 +347,8 @@ static int xgene_gpio_sb_remove(struct platform_device *pdev)
>>               acpi_gpiochip_free_interrupts(&priv->gc);
>>       }
>>
>> +     irq_domain_remove(priv->irq_domain);
>> +
>>       gpiochip_remove(&priv->gc);
>>       return 0;
>>  }
>>
>
> Thanks,
>
>         M.
> --
> Jazz is not dead. It just smells funny...

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [PATCH v5 2/3] Documentation: gpio: Update description for X-Gene standby GPIO controller DTS binding
  2016-02-02  1:46         ` Quan Nguyen
@ 2016-02-12 14:36           ` Rob Herring
  -1 siblings, 0 replies; 24+ messages in thread
From: Rob Herring @ 2016-02-12 14:36 UTC (permalink / raw)
  To: Quan Nguyen
  Cc: Linus Walleij, linux-gpio, devicetree, linux-arm-kernel,
	Thomas Gleixner, Jason Cooper, Marc Zyngier, Y Vo, Phong Vo,
	Loc Ho, Feng Kan, Duc Dang, patches

On Tue, Feb 02, 2016 at 08:46:25AM +0700, Quan Nguyen wrote:
> On Mon, Feb 1, 2016 at 10:35 PM, Rob Herring <robh@kernel.org> wrote:
> > On Fri, Jan 29, 2016 at 11:28:54AM +0700, Quan Nguyen wrote:
> >> Update description for X-Gene standby GPIO controller DTS binding to
> >> support GPIO line configuration as input, output or external IRQ pin.
> >>
> >> Signed-off-by: Y Vo <yvo@apm.com>
> >> Signed-off-by: Quan Nguyen <qnguyen@apm.com>
> >> ---
> >>  .../devicetree/bindings/gpio/gpio-xgene-sb.txt     | 47 ++++++++++++++++++----
> >>  1 file changed, 40 insertions(+), 7 deletions(-)
> >>
> >> diff --git a/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt b/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
> >> index dae1300..7b8b4cb 100644
> >> --- a/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
> >> +++ b/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
> >> @@ -1,10 +1,20 @@
> >>  APM X-Gene Standby GPIO controller bindings
> >>
> >> -This is a gpio controller in the standby domain.
> >> -
> >> -There are 20 GPIO pins from 0..21. There is no GPIO_DS14 or GPIO_DS15,
> >> -only GPIO_DS8..GPIO_DS13 support interrupts. The IRQ mapping
> >> -is currently 1-to-1 on interrupts 0x28 thru 0x2d.
> >> +This is a gpio controller in the standby domain. It also supports interrupt in
> >> +some particular pins which are sourced to its parent interrupt controller
> >> +as diagram below:
> >> +                           +-----------------+
> >> +                           | X-Gene standby  |
> >> +                           | GPIO controller +--------- GPIO_0
> >> ++------------+             |                 | ...
> >> +| Parent IRQ |             |                 +--------- GPIO_8/EXT_INT_0
> >> +| controller |  EXT_INT_0  |                 | ...
> >> +| (GICv2)    +-------------+                 +--------- GPIO_[N+8]/EXT_INT_N
> >> +|            |  ...        |                 |
> >> +|            |  EXT_INT_N  |                 +--------- GPIO_[N+9]
> >> +|            +-------------+                 | ...
> >> +|            |             |                 +--------- GPIO_MAX
> >> ++------------+             +-----------------+
> >>
> >>  Required properties:
> >>  - compatible: "apm,xgene-gpio-sb" for the X-Gene Standby GPIO controller
> >> @@ -15,10 +25,18 @@ Required properties:
> >>               0 = active high
> >>               1 = active low
> >>  - gpio-controller: Marks the device node as a GPIO controller.
> >> -- interrupts: Shall contain exactly 6 interrupts.
> >> +- interrupts: The EXT_INT_0 parent interrupt resource must be listed first.
> >> +- interrupt-parent: Phandle of the parent interrupt controller.
> >> +- interrupt-cells: Should be two.
> >> +       - first cell is 0-N coresponding for EXT_INT_0 to EXT_INT_N.
> >> +       - second cell is used to specify flags.
> >> +- interrupt-controller: Marks the device node as an interrupt controller.
> >> +- apm,nr-gpios: Optional, specify number of gpios pin.
> >> +- apm,nr-irqs: Optional, specify number of interrupt pins.
> >
> > When is this not 6?
> >
> Hi Rob,  by default, this should be 6, but I think this property can
> help in cases:
> + Used only 5(or less) first pin as interrupt.
> + For similar device which has different interrupt pins (8 for example).
> My idea is to make it a bit more generic by using this optional property.

What is similar device? Another SoC? board? Is 6 fixed in the SoC? I 
think you need more specific compatible string with the SoC name in it 
to determine these setting rather than trying to do something generic.

Rob

^ permalink raw reply	[flat|nested] 24+ messages in thread

* [PATCH v5 2/3] Documentation: gpio: Update description for X-Gene standby GPIO controller DTS binding
@ 2016-02-12 14:36           ` Rob Herring
  0 siblings, 0 replies; 24+ messages in thread
From: Rob Herring @ 2016-02-12 14:36 UTC (permalink / raw)
  To: linux-arm-kernel

On Tue, Feb 02, 2016 at 08:46:25AM +0700, Quan Nguyen wrote:
> On Mon, Feb 1, 2016 at 10:35 PM, Rob Herring <robh@kernel.org> wrote:
> > On Fri, Jan 29, 2016 at 11:28:54AM +0700, Quan Nguyen wrote:
> >> Update description for X-Gene standby GPIO controller DTS binding to
> >> support GPIO line configuration as input, output or external IRQ pin.
> >>
> >> Signed-off-by: Y Vo <yvo@apm.com>
> >> Signed-off-by: Quan Nguyen <qnguyen@apm.com>
> >> ---
> >>  .../devicetree/bindings/gpio/gpio-xgene-sb.txt     | 47 ++++++++++++++++++----
> >>  1 file changed, 40 insertions(+), 7 deletions(-)
> >>
> >> diff --git a/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt b/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
> >> index dae1300..7b8b4cb 100644
> >> --- a/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
> >> +++ b/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
> >> @@ -1,10 +1,20 @@
> >>  APM X-Gene Standby GPIO controller bindings
> >>
> >> -This is a gpio controller in the standby domain.
> >> -
> >> -There are 20 GPIO pins from 0..21. There is no GPIO_DS14 or GPIO_DS15,
> >> -only GPIO_DS8..GPIO_DS13 support interrupts. The IRQ mapping
> >> -is currently 1-to-1 on interrupts 0x28 thru 0x2d.
> >> +This is a gpio controller in the standby domain. It also supports interrupt in
> >> +some particular pins which are sourced to its parent interrupt controller
> >> +as diagram below:
> >> +                           +-----------------+
> >> +                           | X-Gene standby  |
> >> +                           | GPIO controller +--------- GPIO_0
> >> ++------------+             |                 | ...
> >> +| Parent IRQ |             |                 +--------- GPIO_8/EXT_INT_0
> >> +| controller |  EXT_INT_0  |                 | ...
> >> +| (GICv2)    +-------------+                 +--------- GPIO_[N+8]/EXT_INT_N
> >> +|            |  ...        |                 |
> >> +|            |  EXT_INT_N  |                 +--------- GPIO_[N+9]
> >> +|            +-------------+                 | ...
> >> +|            |             |                 +--------- GPIO_MAX
> >> ++------------+             +-----------------+
> >>
> >>  Required properties:
> >>  - compatible: "apm,xgene-gpio-sb" for the X-Gene Standby GPIO controller
> >> @@ -15,10 +25,18 @@ Required properties:
> >>               0 = active high
> >>               1 = active low
> >>  - gpio-controller: Marks the device node as a GPIO controller.
> >> -- interrupts: Shall contain exactly 6 interrupts.
> >> +- interrupts: The EXT_INT_0 parent interrupt resource must be listed first.
> >> +- interrupt-parent: Phandle of the parent interrupt controller.
> >> +- interrupt-cells: Should be two.
> >> +       - first cell is 0-N coresponding for EXT_INT_0 to EXT_INT_N.
> >> +       - second cell is used to specify flags.
> >> +- interrupt-controller: Marks the device node as an interrupt controller.
> >> +- apm,nr-gpios: Optional, specify number of gpios pin.
> >> +- apm,nr-irqs: Optional, specify number of interrupt pins.
> >
> > When is this not 6?
> >
> Hi Rob,  by default, this should be 6, but I think this property can
> help in cases:
> + Used only 5(or less) first pin as interrupt.
> + For similar device which has different interrupt pins (8 for example).
> My idea is to make it a bit more generic by using this optional property.

What is similar device? Another SoC? board? Is 6 fixed in the SoC? I 
think you need more specific compatible string with the SoC name in it 
to determine these setting rather than trying to do something generic.

Rob

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [PATCH v5 2/3] Documentation: gpio: Update description for X-Gene standby GPIO controller DTS binding
  2016-02-12 14:36           ` Rob Herring
@ 2016-02-13  3:21             ` Quan Nguyen
  -1 siblings, 0 replies; 24+ messages in thread
From: Quan Nguyen @ 2016-02-13  3:21 UTC (permalink / raw)
  To: Rob Herring
  Cc: Linus Walleij, linux-gpio, devicetree, linux-arm-kernel,
	Thomas Gleixner, Jason Cooper, Marc Zyngier, Y Vo, Phong Vo,
	Loc Ho, Feng Kan, Duc Dang, patches

On Fri, Feb 12, 2016 at 9:36 PM, Rob Herring <robh@kernel.org> wrote:
> On Tue, Feb 02, 2016 at 08:46:25AM +0700, Quan Nguyen wrote:
>> On Mon, Feb 1, 2016 at 10:35 PM, Rob Herring <robh@kernel.org> wrote:
>> > On Fri, Jan 29, 2016 at 11:28:54AM +0700, Quan Nguyen wrote:
>> >> Update description for X-Gene standby GPIO controller DTS binding to
>> >> support GPIO line configuration as input, output or external IRQ pin.
>> >>
>> >> Signed-off-by: Y Vo <yvo@apm.com>
>> >> Signed-off-by: Quan Nguyen <qnguyen@apm.com>
>> >> ---
>> >>  .../devicetree/bindings/gpio/gpio-xgene-sb.txt     | 47 ++++++++++++++++++----
>> >>  1 file changed, 40 insertions(+), 7 deletions(-)
>> >>
>> >> diff --git a/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt b/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
>> >> index dae1300..7b8b4cb 100644
>> >> --- a/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
>> >> +++ b/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
>> >> @@ -1,10 +1,20 @@
>> >>  APM X-Gene Standby GPIO controller bindings
>> >>
>> >> -This is a gpio controller in the standby domain.
>> >> -
>> >> -There are 20 GPIO pins from 0..21. There is no GPIO_DS14 or GPIO_DS15,
>> >> -only GPIO_DS8..GPIO_DS13 support interrupts. The IRQ mapping
>> >> -is currently 1-to-1 on interrupts 0x28 thru 0x2d.
>> >> +This is a gpio controller in the standby domain. It also supports interrupt in
>> >> +some particular pins which are sourced to its parent interrupt controller
>> >> +as diagram below:
>> >> +                           +-----------------+
>> >> +                           | X-Gene standby  |
>> >> +                           | GPIO controller +--------- GPIO_0
>> >> ++------------+             |                 | ...
>> >> +| Parent IRQ |             |                 +--------- GPIO_8/EXT_INT_0
>> >> +| controller |  EXT_INT_0  |                 | ...
>> >> +| (GICv2)    +-------------+                 +--------- GPIO_[N+8]/EXT_INT_N
>> >> +|            |  ...        |                 |
>> >> +|            |  EXT_INT_N  |                 +--------- GPIO_[N+9]
>> >> +|            +-------------+                 | ...
>> >> +|            |             |                 +--------- GPIO_MAX
>> >> ++------------+             +-----------------+
>> >>
>> >>  Required properties:
>> >>  - compatible: "apm,xgene-gpio-sb" for the X-Gene Standby GPIO controller
>> >> @@ -15,10 +25,18 @@ Required properties:
>> >>               0 = active high
>> >>               1 = active low
>> >>  - gpio-controller: Marks the device node as a GPIO controller.
>> >> -- interrupts: Shall contain exactly 6 interrupts.
>> >> +- interrupts: The EXT_INT_0 parent interrupt resource must be listed first.
>> >> +- interrupt-parent: Phandle of the parent interrupt controller.
>> >> +- interrupt-cells: Should be two.
>> >> +       - first cell is 0-N coresponding for EXT_INT_0 to EXT_INT_N.
>> >> +       - second cell is used to specify flags.
>> >> +- interrupt-controller: Marks the device node as an interrupt controller.
>> >> +- apm,nr-gpios: Optional, specify number of gpios pin.
>> >> +- apm,nr-irqs: Optional, specify number of interrupt pins.
>> >
>> > When is this not 6?
>> >
>> Hi Rob,  by default, this should be 6, but I think this property can
>> help in cases:
>> + Used only 5(or less) first pin as interrupt.
>> + For similar device which has different interrupt pins (8 for example).
>> My idea is to make it a bit more generic by using this optional property.
>
> What is similar device? Another SoC? board? Is 6 fixed in the SoC? I
> think you need more specific compatible string with the SoC name in it
> to determine these setting rather than trying to do something generic.
>

I agree, Rob, compatible string should be suit for other SoC, but I think
we still need this property for case of less than 6 interrupts are needed.

Thanks,
-- Quan Nguyen

^ permalink raw reply	[flat|nested] 24+ messages in thread

* [PATCH v5 2/3] Documentation: gpio: Update description for X-Gene standby GPIO controller DTS binding
@ 2016-02-13  3:21             ` Quan Nguyen
  0 siblings, 0 replies; 24+ messages in thread
From: Quan Nguyen @ 2016-02-13  3:21 UTC (permalink / raw)
  To: linux-arm-kernel

On Fri, Feb 12, 2016 at 9:36 PM, Rob Herring <robh@kernel.org> wrote:
> On Tue, Feb 02, 2016 at 08:46:25AM +0700, Quan Nguyen wrote:
>> On Mon, Feb 1, 2016 at 10:35 PM, Rob Herring <robh@kernel.org> wrote:
>> > On Fri, Jan 29, 2016 at 11:28:54AM +0700, Quan Nguyen wrote:
>> >> Update description for X-Gene standby GPIO controller DTS binding to
>> >> support GPIO line configuration as input, output or external IRQ pin.
>> >>
>> >> Signed-off-by: Y Vo <yvo@apm.com>
>> >> Signed-off-by: Quan Nguyen <qnguyen@apm.com>
>> >> ---
>> >>  .../devicetree/bindings/gpio/gpio-xgene-sb.txt     | 47 ++++++++++++++++++----
>> >>  1 file changed, 40 insertions(+), 7 deletions(-)
>> >>
>> >> diff --git a/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt b/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
>> >> index dae1300..7b8b4cb 100644
>> >> --- a/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
>> >> +++ b/Documentation/devicetree/bindings/gpio/gpio-xgene-sb.txt
>> >> @@ -1,10 +1,20 @@
>> >>  APM X-Gene Standby GPIO controller bindings
>> >>
>> >> -This is a gpio controller in the standby domain.
>> >> -
>> >> -There are 20 GPIO pins from 0..21. There is no GPIO_DS14 or GPIO_DS15,
>> >> -only GPIO_DS8..GPIO_DS13 support interrupts. The IRQ mapping
>> >> -is currently 1-to-1 on interrupts 0x28 thru 0x2d.
>> >> +This is a gpio controller in the standby domain. It also supports interrupt in
>> >> +some particular pins which are sourced to its parent interrupt controller
>> >> +as diagram below:
>> >> +                           +-----------------+
>> >> +                           | X-Gene standby  |
>> >> +                           | GPIO controller +--------- GPIO_0
>> >> ++------------+             |                 | ...
>> >> +| Parent IRQ |             |                 +--------- GPIO_8/EXT_INT_0
>> >> +| controller |  EXT_INT_0  |                 | ...
>> >> +| (GICv2)    +-------------+                 +--------- GPIO_[N+8]/EXT_INT_N
>> >> +|            |  ...        |                 |
>> >> +|            |  EXT_INT_N  |                 +--------- GPIO_[N+9]
>> >> +|            +-------------+                 | ...
>> >> +|            |             |                 +--------- GPIO_MAX
>> >> ++------------+             +-----------------+
>> >>
>> >>  Required properties:
>> >>  - compatible: "apm,xgene-gpio-sb" for the X-Gene Standby GPIO controller
>> >> @@ -15,10 +25,18 @@ Required properties:
>> >>               0 = active high
>> >>               1 = active low
>> >>  - gpio-controller: Marks the device node as a GPIO controller.
>> >> -- interrupts: Shall contain exactly 6 interrupts.
>> >> +- interrupts: The EXT_INT_0 parent interrupt resource must be listed first.
>> >> +- interrupt-parent: Phandle of the parent interrupt controller.
>> >> +- interrupt-cells: Should be two.
>> >> +       - first cell is 0-N coresponding for EXT_INT_0 to EXT_INT_N.
>> >> +       - second cell is used to specify flags.
>> >> +- interrupt-controller: Marks the device node as an interrupt controller.
>> >> +- apm,nr-gpios: Optional, specify number of gpios pin.
>> >> +- apm,nr-irqs: Optional, specify number of interrupt pins.
>> >
>> > When is this not 6?
>> >
>> Hi Rob,  by default, this should be 6, but I think this property can
>> help in cases:
>> + Used only 5(or less) first pin as interrupt.
>> + For similar device which has different interrupt pins (8 for example).
>> My idea is to make it a bit more generic by using this optional property.
>
> What is similar device? Another SoC? board? Is 6 fixed in the SoC? I
> think you need more specific compatible string with the SoC name in it
> to determine these setting rather than trying to do something generic.
>

I agree, Rob, compatible string should be suit for other SoC, but I think
we still need this property for case of less than 6 interrupts are needed.

Thanks,
-- Quan Nguyen

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [PATCH v5 1/3] gpio: xgene: Enable X-Gene standby GPIO as interrupt controller
  2016-02-10 15:09     ` Marc Zyngier
@ 2016-02-15 23:26       ` Linus Walleij
  -1 siblings, 0 replies; 24+ messages in thread
From: Linus Walleij @ 2016-02-15 23:26 UTC (permalink / raw)
  To: Marc Zyngier
  Cc: Quan Nguyen, linux-gpio, devicetree, linux-arm-kernel,
	Thomas Gleixner, Jason Cooper, Y Vo, Phong Vo, Loc Ho, Feng Kan,
	Duc Dang, patches

On Wed, Feb 10, 2016 at 4:09 PM, Marc Zyngier <marc.zyngier@arm.com> wrote:
> On 29/01/16 04:28, Quan Nguyen wrote:

>> +static void xgene_gpio_sb_domain_deactivate(struct irq_domain *d,
>> +             struct irq_data *irq_data)
>> +{
>> +     struct xgene_gpio_sb *priv = d->host_data;
>> +     u32 gpio = HWIRQ_TO_GPIO(priv, irq_data->hwirq);
>> +
>> +     gpiochip_unlock_as_irq(&priv->gc, gpio);
>
> Is it right to do the unlock both in irq_shutdown and domain_deactivate?
> This seems a bit odd to me to have such an inbalance. My hunch is that
> you should either implement irq_startup, do the locking there and drop
> the unlock drop deactivate, or kill irq_shutdown.
>
> Linus, what do you think?

Those functions should be called in .irq_alloc/release_resources(), that is
why Gleixner added those callbacks in the first place.

Yours,
Linus Walleij

^ permalink raw reply	[flat|nested] 24+ messages in thread

* [PATCH v5 1/3] gpio: xgene: Enable X-Gene standby GPIO as interrupt controller
@ 2016-02-15 23:26       ` Linus Walleij
  0 siblings, 0 replies; 24+ messages in thread
From: Linus Walleij @ 2016-02-15 23:26 UTC (permalink / raw)
  To: linux-arm-kernel

On Wed, Feb 10, 2016 at 4:09 PM, Marc Zyngier <marc.zyngier@arm.com> wrote:
> On 29/01/16 04:28, Quan Nguyen wrote:

>> +static void xgene_gpio_sb_domain_deactivate(struct irq_domain *d,
>> +             struct irq_data *irq_data)
>> +{
>> +     struct xgene_gpio_sb *priv = d->host_data;
>> +     u32 gpio = HWIRQ_TO_GPIO(priv, irq_data->hwirq);
>> +
>> +     gpiochip_unlock_as_irq(&priv->gc, gpio);
>
> Is it right to do the unlock both in irq_shutdown and domain_deactivate?
> This seems a bit odd to me to have such an inbalance. My hunch is that
> you should either implement irq_startup, do the locking there and drop
> the unlock drop deactivate, or kill irq_shutdown.
>
> Linus, what do you think?

Those functions should be called in .irq_alloc/release_resources(), that is
why Gleixner added those callbacks in the first place.

Yours,
Linus Walleij

^ permalink raw reply	[flat|nested] 24+ messages in thread

end of thread, other threads:[~2016-02-15 23:26 UTC | newest]

Thread overview: 24+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-01-29  4:28 [PATCH v5 0/3] Enable X-Gene standby GPIO as interrupt controller Quan Nguyen
2016-01-29  4:28 ` Quan Nguyen
2016-01-29  4:28 ` [PATCH v5 1/3] gpio: xgene: " Quan Nguyen
2016-01-29  4:28   ` Quan Nguyen
     [not found]   ` <1454041735-8434-2-git-send-email-qnguyen-qTEPVZfXA3Y@public.gmane.org>
2016-02-10 10:58     ` Linus Walleij
2016-02-10 10:58       ` Linus Walleij
2016-02-10 15:09   ` Marc Zyngier
2016-02-10 15:09     ` Marc Zyngier
     [not found]     ` <56BB52B8.4040501-5wv7dgnIgG8@public.gmane.org>
2016-02-11 16:35       ` Quan Nguyen
2016-02-11 16:35         ` Quan Nguyen
2016-02-15 23:26     ` Linus Walleij
2016-02-15 23:26       ` Linus Walleij
2016-01-29  4:28 ` [PATCH v5 2/3] Documentation: gpio: Update description for X-Gene standby GPIO controller DTS binding Quan Nguyen
2016-01-29  4:28   ` Quan Nguyen
     [not found]   ` <1454041735-8434-3-git-send-email-qnguyen-qTEPVZfXA3Y@public.gmane.org>
2016-02-01 15:35     ` Rob Herring
2016-02-01 15:35       ` Rob Herring
2016-02-02  1:46       ` Quan Nguyen
2016-02-02  1:46         ` Quan Nguyen
2016-02-12 14:36         ` Rob Herring
2016-02-12 14:36           ` Rob Herring
2016-02-13  3:21           ` Quan Nguyen
2016-02-13  3:21             ` Quan Nguyen
2016-01-29  4:28 ` [PATCH v5 3/3] arm64: dts: Update X-Gene standby GPIO controller DTS entries Quan Nguyen
2016-01-29  4:28   ` Quan Nguyen

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.