[PATCH 1/2] powerpc: ipic: use set_irq_chip to ensure irq_chip defaults are applied

[PATCH 1/2] powerpc: ipic: use set_irq_chip to ensure irq_chip defaults are applied

[Esben Haabendal]

Signed-off-by: Esben Haabendal <eha@doredevelopment.dk>
---
 arch/powerpc/sysdev/ipic.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/powerpc/sysdev/ipic.c b/arch/powerpc/sysdev/ipic.c
index d7b9b9c..8464b86 100644
--- a/arch/powerpc/sysdev/ipic.c
+++ b/arch/powerpc/sysdev/ipic.c
@@ -630,10 +630,10 @@ static int ipic_set_irq_type(unsigned int virq, unsigned int flow_type)
 	if (flow_type & IRQ_TYPE_LEVEL_LOW)  {
 		desc->status |= IRQ_LEVEL;
 		desc->handle_irq = handle_level_irq;
-		desc->chip = &ipic_level_irq_chip;
+		set_irq_chip(virq, &ipic_level_irq_chip);
 	} else {
 		desc->handle_irq = handle_edge_irq;
-		desc->chip = &ipic_edge_irq_chip;
+		set_irq_chip(virq, &ipic_edge_irq_chip);
 	}
 
 	/* only EXT IRQ senses are programmable on ipic
-- 
1.7.1

Re: [PATCH 1/2] powerpc: ipic: use set_irq_chip to ensure irq_chip defaults are applied

[Thomas Gleixner]

On Fri, 4 Jun 2010, Esben Haabendal wrote:

What's the rationale of this patch and which problem does it solve ?

> Signed-off-by: Esben Haabendal <eha@doredevelopment.dk>
> ---
>  arch/powerpc/sysdev/ipic.c |    4 ++--
>  1 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/powerpc/sysdev/ipic.c b/arch/powerpc/sysdev/ipic.c
> index d7b9b9c..8464b86 100644
> --- a/arch/powerpc/sysdev/ipic.c
> +++ b/arch/powerpc/sysdev/ipic.c
> @@ -630,10 +630,10 @@ static int ipic_set_irq_type(unsigned int virq, unsigned int flow_type)
>  	if (flow_type & IRQ_TYPE_LEVEL_LOW)  {
>  		desc->status |= IRQ_LEVEL;
>  		desc->handle_irq = handle_level_irq;
> -		desc->chip = &ipic_level_irq_chip;
> +		set_irq_chip(virq, &ipic_level_irq_chip);

This patch has never been tested with spinlock debugging enabled and
will break SMP as it causes a deadlock on irq_desc->lock.

Again: See Documentation/Submit*

Thanks,

	tglx

Re: [PATCH 1/2] powerpc: ipic: use set_irq_chip to ensure irq_chip defaults are applied

[Esben Haabendal]

On Mon, 2010-06-07 at 01:45 +0200, Thomas Gleixner wrote:

> This patch has never been tested with spinlock debugging enabled and
> will break SMP as it causes a deadlock on irq_desc->lock.
> 
> Again: See Documentation/Submit*

Ok, will do so.

But do you see the problem?  In __setup_irq(), irq_chip_set_defaults()
is called first, and then __irq_set_trigger(), which calls
chip->set_type(). When you request an edge irq with ipic, you get the
defaults applied to ipic_level_irq_chip, and then it sets
ipic_edge_irq_chip, which now is missing startup(), enable(), and so
on.

Would it be better to change the call order in __setup_irq(), and
call irq_chip_set_defaults after __irq_set_trigger() ?   Or perhaps
even calling it twice (again after __irq_set_trigger()) ?

/Esben

Re: [PATCH 1/2] powerpc: ipic: use set_irq_chip to ensure irq_chip defaults are applied

[Thomas Gleixner]

On Mon, 7 Jun 2010, Esben Haabendal wrote:

> On Mon, 2010-06-07 at 01:45 +0200, Thomas Gleixner wrote:
> 
> > This patch has never been tested with spinlock debugging enabled and
> > will break SMP as it causes a deadlock on irq_desc->lock.
> > 
> > Again: See Documentation/Submit*
> 
> Ok, will do so.
> 
> But do you see the problem?  In __setup_irq(), irq_chip_set_defaults()
> is called first, and then __irq_set_trigger(), which calls
> chip->set_type(). When you request an edge irq with ipic, you get the
> defaults applied to ipic_level_irq_chip, and then it sets
> ipic_edge_irq_chip, which now is missing startup(), enable(), and so
> on.
> 
> Would it be better to change the call order in __setup_irq(), and
> call irq_chip_set_defaults after __irq_set_trigger() ?   Or perhaps
> even calling it twice (again after __irq_set_trigger()) ?

Grmpf, set_type() was never meant to change the chip. It's fatal do
so, as the code in setup_irq already has a reference to desc->chip and
calls the wrong functions anyway aside of having not run through the
set defaults code.

That needs more thought, as it requires to reload the reference.

Ben, any thoughts ?

Thanks,

	tglx

Re: [PATCH 1/2] powerpc: ipic: use set_irq_chip to ensure irq_chip defaults are applied

[Benjamin Herrenschmidt]

On Mon, 2010-06-07 at 12:22 +0200, Thomas Gleixner wrote:

> > Would it be better to change the call order in __setup_irq(), and
> > call irq_chip_set_defaults after __irq_set_trigger() ?   Or perhaps
> > even calling it twice (again after __irq_set_trigger()) ?
> 
> Grmpf, set_type() was never meant to change the chip. It's fatal do
> so, as the code in setup_irq already has a reference to desc->chip and
> calls the wrong functions anyway aside of having not run through the
> set defaults code.
> 
> That needs more thought, as it requires to reload the reference.
> 
> Ben, any thoughts ?

Not off hand. It's common to have a different chip ops for different
types, so we need to sort that out. I'll try to have a look asap but I'm
really really swamped at the moment, so I can make no promise that I'll
come up with a solution tomorrow.

/me off to bed

Cheers,
Ben.

[PATCH 2/2] gpio: pca953x: add powerpc irq support

[Esben Haabendal]

Signed-off-by: Esben Haabendal <eha@doredevelopment.dk>
---
 drivers/gpio/pca953x.c |  347 +++++++++++++++++++++++++++++++++---------------
 1 files changed, 240 insertions(+), 107 deletions(-)

diff --git a/drivers/gpio/pca953x.c b/drivers/gpio/pca953x.c
index a2b12aa..a47d55f 100644
--- a/drivers/gpio/pca953x.c
+++ b/drivers/gpio/pca953x.c
@@ -3,6 +3,7 @@
  *
  *  Copyright (C) 2005 Ben Gardner <bgardner@wabtec.com>
  *  Copyright (C) 2007 Marvell International Ltd.
+ *  Copyright (C) 2010 DoreDevelopment ApS
  *
  *  Derived from drivers/i2c/chips/pca9539.c
  *
@@ -67,8 +68,11 @@ struct pca953x_chip {
 	uint16_t irq_stat;
 	uint16_t irq_trig_raise;
 	uint16_t irq_trig_fall;
-	int	 irq_base;
-#endif
+	int irq_base;
+#ifdef CONFIG_PPC
+	struct irq_host *irq_host;
+#endif /* CONFIG_PPC */
+#endif /* CONFIG_GPIO_PCA953X_IRQ */
 
 	struct i2c_client *client;
 	struct pca953x_platform_data *dyn_pdata;
@@ -120,6 +124,10 @@ static int pca953x_gpio_direction_input(struct gpio_chip *gc, unsigned off)
 	chip = container_of(gc, struct pca953x_chip, gpio_chip);
 
 	reg_val = chip->reg_direction | (1u << off);
+
+	if (reg_val == chip->reg_direction)
+		return 0;
+
 	ret = pca953x_write_reg(chip, PCA953X_DIRECTION, reg_val);
 	if (ret)
 		return ret;
@@ -220,63 +228,41 @@ static void pca953x_setup_gpio(struct pca953x_chip *chip, int gpios)
 }
 
 #ifdef CONFIG_GPIO_PCA953X_IRQ
-static int pca953x_gpio_to_irq(struct gpio_chip *gc, unsigned off)
-{
-	struct pca953x_chip *chip;
 
-	chip = container_of(gc, struct pca953x_chip, gpio_chip);
-	return chip->irq_base + off;
-}
+#ifdef CONFIG_PPC
 
-static void pca953x_irq_mask(unsigned int irq)
-{
-	struct pca953x_chip *chip = get_irq_chip_data(irq);
-
-	chip->irq_mask &= ~(1 << (irq - chip->irq_base));
-}
+#define pca953x_pin_to_virq(chip, pin) \
+	(irq_linear_revmap((chip)->irq_host, pin))
+#define pca953x_virq_to_pin(chip, virq) \
+	(virq_to_hw(virq))
 
-static void pca953x_irq_unmask(unsigned int irq)
+static void pca953x_irq_mask(unsigned int virq)
 {
-	struct pca953x_chip *chip = get_irq_chip_data(irq);
-
-	chip->irq_mask |= 1 << (irq - chip->irq_base);
+	struct pca953x_chip *chip = get_irq_chip_data(virq);
+	int pin = pca953x_virq_to_pin(chip, virq);
+	pr_debug("%s: virq=%d pin=%d\n", __func__, virq, pin);
+	chip->irq_mask &= ~(1 << pin);
 }
 
-static void pca953x_irq_bus_lock(unsigned int irq)
+static void pca953x_irq_unmask(unsigned int virq)
 {
-	struct pca953x_chip *chip = get_irq_chip_data(irq);
-
-	mutex_lock(&chip->irq_lock);
+	struct pca953x_chip *chip = get_irq_chip_data(virq);
+	int pin = pca953x_virq_to_pin(chip, virq);
+	pr_debug("%s: virq=%d hwirq=%d\n", __func__, virq, pin);
+	chip->irq_mask |= 1 << pin;
 }
 
-static void pca953x_irq_bus_sync_unlock(unsigned int irq)
+static int pca953x_irq_set_type(unsigned int virq, unsigned int type)
 {
-	struct pca953x_chip *chip = get_irq_chip_data(irq);
-	uint16_t new_irqs;
-	uint16_t level;
+	struct pca953x_chip *chip = get_irq_chip_data(virq);
+	int pin = pca953x_virq_to_pin(chip, virq);
+	uint16_t mask = 1 << pin;
 
-	/* Look for any newly setup interrupt */
-	new_irqs = chip->irq_trig_fall | chip->irq_trig_raise;
-	new_irqs &= ~chip->reg_direction;
-
-	while (new_irqs) {
-		level = __ffs(new_irqs);
-		pca953x_gpio_direction_input(&chip->gpio_chip, level);
-		new_irqs &= ~(1 << level);
-	}
-
-	mutex_unlock(&chip->irq_lock);
-}
-
-static int pca953x_irq_set_type(unsigned int irq, unsigned int type)
-{
-	struct pca953x_chip *chip = get_irq_chip_data(irq);
-	uint16_t level = irq - chip->irq_base;
-	uint16_t mask = 1 << level;
+	pr_debug("%s: virq=%d pin=%d type=0x%x\n", __func__, virq, pin, type);
 
 	if (!(type & IRQ_TYPE_EDGE_BOTH)) {
-		dev_err(&chip->client->dev, "irq %d: unsupported type %d\n",
-			irq, type);
+		dev_err(&chip->client->dev, "virq %d: unsupported type %d\n",
+			virq, type);
 		return -EINVAL;
 	}
 
@@ -297,11 +283,140 @@ static struct irq_chip pca953x_irq_chip = {
 	.name			= "pca953x",
 	.mask			= pca953x_irq_mask,
 	.unmask			= pca953x_irq_unmask,
+#ifndef CONFIG_PPC
 	.bus_lock		= pca953x_irq_bus_lock,
 	.bus_sync_unlock	= pca953x_irq_bus_sync_unlock,
+#endif /* !CONFIG_PPC */
 	.set_type		= pca953x_irq_set_type,
 };
 
+static int pca953x_irq_host_match(struct irq_host *h, struct device_node *node)
+{
+	pr_debug("%s: node=%s\n", __func__, node->full_name);
+	return h->of_node == NULL || h->of_node == node;
+}
+
+static int pca953x_irq_host_map(struct irq_host *h, unsigned int virq,
+				irq_hw_number_t hwirq)
+{
+	struct pca953x_chip *chip = h->host_data;
+
+	pr_debug("%s: virq=%d hwirq=%ld\n", __func__, virq, hwirq);
+
+	set_irq_chip_data(virq, chip);
+	set_irq_chip_and_handler(virq, &pca953x_irq_chip, handle_edge_irq);
+	set_irq_nested_thread(virq, 1);
+
+	/* On powerpc, we set the direction to input on map(),
+	 * avoiding the use of genirq buslock, and thereby not
+	 * restricting the use of generic interrupt management
+	 * functions. */
+	pca953x_gpio_direction_input(&chip->gpio_chip, hwirq);
+
+	return 0;
+}
+
+static int pca953x_irq_host_xlate(struct irq_host *h, struct device_node *node,
+				  const u32 *intspec, unsigned int intsize,
+				  irq_hw_number_t *hwirq, unsigned int *flags)
+{
+	pr_debug("%s: irq_host=%p node=%s\n", __func__, h, node->full_name);
+	*hwirq = intspec[0];
+	if (intsize > 1)
+		*flags = intspec[1];
+	else
+		*flags = IRQ_TYPE_EDGE_BOTH;
+	pr_debug("%s: hwirq=%ld flags=0x%x\n", __func__, *hwirq, *flags);
+	return 0;
+}
+
+static struct irq_host_ops pca953x_irq_host_ops = {
+	.match	= pca953x_irq_host_match,
+	.map	= pca953x_irq_host_map,
+	.xlate	= pca953x_irq_host_xlate,
+};
+
+#define pca953x_irq_setup_premap(chip) do {} while (0)
+
+static void pca953x_irq_alloc_host(struct pca953x_chip *chip,
+				   struct device_node *node)
+{
+	chip->irq_host = irq_alloc_host(
+		node, IRQ_HOST_MAP_LINEAR, chip->gpio_chip.ngpio,
+		&pca953x_irq_host_ops, -1);
+	chip->irq_host->host_data = chip;
+}
+
+#else /* CONFIG_PPC */
+
+#define pca953x_pin_to_virq(chip, pin) \
+	((chip)->irq_base + pin)
+#define pca953x_virq_to_pin(chip, virq) \
+	(virq - (chip)->irq_base)
+
+/* Use the genirq buslock support on non-powerpc archs for setting
+ * gpio direction on enable_irq, which unfortunately means that some
+ * of the generic interrupt management functions like enable_irq,
+ * disable_irq and so on cannot be used in atomic context on irq's
+ * from this interrupt controller. */
+static void pca953x_irq_bus_lock(unsigned int virq)
+{
+	struct pca953x_chip *chip = get_irq_chip_data(virq);
+	pr_debug("%s: virq=%d\n", __func__, virq);
+	mutex_lock(&chip->irq_lock);
+}
+
+static void pca953x_irq_bus_sync_unlock(unsigned int virq)
+{
+	struct pca953x_chip *chip = get_irq_chip_data(virq);
+	uint16_t new_irqs;
+	uint16_t pin;
+	pr_debug("%s: virq=%d\n", __func__, virq);
+
+	/* Look for any newly setup interrupt */
+	new_irqs = chip->irq_trig_fall | chip->irq_trig_raise;
+	new_irqs &= ~chip->reg_direction;
+
+	while (new_irqs) {
+		pin = __ffs(new_irqs);
+		pca953x_gpio_direction_input(&chip->gpio_chip, pin);
+		new_irqs &= ~(1 << pin);
+	}
+
+	mutex_unlock(&chip->irq_lock);
+}
+
+static void pca953x_irq_setup_premap(struct pca953x_chipo *chip)
+{
+	int pin;
+
+	for (pin = 0; pin < chip->gpio_chip.ngpio; pin++) {
+		int irq = pin + chip->irq_base;
+
+		//irq_to_desc_alloc_node(irq, 0);
+		set_irq_chip_data(irq, chip);
+		set_irq_chip_and_handler(irq, &pca953x_irq_chip,
+					 handle_edge_irq);
+		set_irq_nested_thread(irq, 1);
+#ifdef CONFIG_ARM
+		set_irq_flags(irq, IRQF_VALID);
+#else /* CONFIG_ARM */
+		set_irq_noprobe(irq);
+#endif /* CONFIG_ARM */
+	}
+}
+
+#define pca953x_irq_alloc_host(chip) do {} while (0)
+
+#endif /* CONFIG_PPC */
+
+static int pca953x_gpio_to_irq(struct gpio_chip *gc, unsigned off)
+{
+	struct pca953x_chip *chip;
+	chip = container_of(gc, struct pca953x_chip, gpio_chip);
+	return pca953x_pin_to_virq(chip, off);
+}
+
 static uint16_t pca953x_irq_pending(struct pca953x_chip *chip)
 {
 	uint16_t cur_stat;
@@ -317,14 +432,20 @@ static uint16_t pca953x_irq_pending(struct pca953x_chip *chip)
 	/* Remove output pins from the equation */
 	cur_stat &= chip->reg_direction;
 
+	/* Remember the input */
 	old_stat = chip->irq_stat;
+	chip->irq_stat = cur_stat;
+
+	/* Find changed inputs which are not masked out */
 	trigger = (cur_stat ^ old_stat) & chip->irq_mask;
 
+	pr_debug("%s: irq_mask=0x%04x old_stat=0x%04x cur_stat=0x%04x\n",
+		 __func__, chip->irq_mask, old_stat, cur_stat);
+
 	if (!trigger)
 		return 0;
 
-	chip->irq_stat = cur_stat;
-
+	/* And finally, get the pending irq's */
 	pending = (old_stat & chip->irq_trig_fall) |
 		  (cur_stat & chip->irq_trig_raise);
 	pending &= trigger;
@@ -332,22 +453,26 @@ static uint16_t pca953x_irq_pending(struct pca953x_chip *chip)
 	return pending;
 }
 
-static irqreturn_t pca953x_irq_handler(int irq, void *devid)
+static irqreturn_t pca953x_irq_thread_fn(int irq, void *devid)
 {
 	struct pca953x_chip *chip = devid;
 	uint16_t pending;
-	uint16_t level;
+	uint16_t pin;
 
 	pending = pca953x_irq_pending(chip);
 
+	pr_debug("%s: virq=%d pending=0x%04x\n", __func__, irq, pending);
+
 	if (!pending)
 		return IRQ_HANDLED;
 
 	do {
-		level = __ffs(pending);
-		handle_nested_irq(level + chip->irq_base);
-
-		pending &= ~(1 << level);
+		unsigned int gpio_irq;
+		pin = __ffs(pending);
+		gpio_irq = pca953x_pin_to_virq(chip, pin);
+		pr_debug("%s: pin=%hu gpio_irq=%d\n", __func__, pin, gpio_irq);
+		handle_nested_irq(gpio_irq);
+		pending &= ~(1 << pin);
 	} while (pending);
 
 	return IRQ_HANDLED;
@@ -360,51 +485,39 @@ static int pca953x_irq_setup(struct pca953x_chip *chip,
 	struct pca953x_platform_data *pdata = client->dev.platform_data;
 	int ret;
 
-	if (pdata->irq_base && (id->driver_data & PCA953X_INT)) {
-		int lvl;
+	if (! ((pdata->irq_base || chip->irq_host) &&
+	       (id->driver_data & PCA953X_INT)))
+		return 0;
 
-		ret = pca953x_read_reg(chip, PCA953X_INPUT,
-				       &chip->irq_stat);
-		if (ret)
-			goto out_failed;
+	ret = pca953x_read_reg(chip, PCA953X_INPUT, &chip->irq_stat);
+	if (ret)
+		goto out_failed;
 
-		/*
-		 * There is no way to know which GPIO line generated the
-		 * interrupt.  We have to rely on the previous read for
-		 * this purpose.
-		 */
-		chip->irq_stat &= chip->reg_direction;
-		chip->irq_base = pdata->irq_base;
-		mutex_init(&chip->irq_lock);
+	/*
+	 * There is no way to know which GPIO line generated the
+	 * interrupt.  We have to rely on the previous read for
+	 * this purpose.
+	 */
+	chip->irq_stat &= chip->reg_direction;
+	mutex_init(&chip->irq_lock);
 
-		for (lvl = 0; lvl < chip->gpio_chip.ngpio; lvl++) {
-			int irq = lvl + chip->irq_base;
+	chip->irq_base = pdata->irq_base;
 
-			set_irq_chip_data(irq, chip);
-			set_irq_chip_and_handler(irq, &pca953x_irq_chip,
-						 handle_edge_irq);
-			set_irq_nested_thread(irq, 1);
-#ifdef CONFIG_ARM
-			set_irq_flags(irq, IRQF_VALID);
-#else
-			set_irq_noprobe(irq);
-#endif
-		}
-
-		ret = request_threaded_irq(client->irq,
-					   NULL,
-					   pca953x_irq_handler,
-					   IRQF_TRIGGER_FALLING | IRQF_ONESHOT,
-					   dev_name(&client->dev), chip);
-		if (ret) {
-			dev_err(&client->dev, "failed to request irq %d\n",
-				client->irq);
-			goto out_failed;
-		}
+	pca953x_irq_setup_premap(chip);
 
-		chip->gpio_chip.to_irq = pca953x_gpio_to_irq;
+	pr_debug("%s: requesting irq=%d\n", __func__, client->irq);
+	ret = request_threaded_irq(client->irq,
+				   NULL, pca953x_irq_thread_fn,
+				   IRQF_TRIGGER_FALLING | IRQF_ONESHOT,
+				   dev_name(&client->dev), chip);
+	if (ret) {
+		dev_err(&client->dev, "failed to request irq %d\n",
+			client->irq);
+		goto out_failed;
 	}
 
+	chip->gpio_chip.to_irq = pca953x_gpio_to_irq;
+
 	return 0;
 
 out_failed:
@@ -414,17 +527,20 @@ out_failed:
 
 static void pca953x_irq_teardown(struct pca953x_chip *chip)
 {
-	if (chip->irq_base)
+	if (chip->irq_host || chip->irq_base)
 		free_irq(chip->client->irq, chip);
 }
+
 #else /* CONFIG_GPIO_PCA953X_IRQ */
+
 static int pca953x_irq_setup(struct pca953x_chip *chip,
 			     const struct i2c_device_id *id)
 {
 	struct i2c_client *client = chip->client;
 	struct pca953x_platform_data *pdata = client->dev.platform_data;
 
-	if (pdata->irq_base && (id->driver_data & PCA953X_INT))
+	if ((chip->irq_host || pdata->irq_base) &&
+	    (id->driver_data & PCA953X_INT))
 		dev_warn(&client->dev, "interrupt support not compiled in\n");
 
 	return 0;
@@ -433,21 +549,21 @@ static int pca953x_irq_setup(struct pca953x_chip *chip,
 static void pca953x_irq_teardown(struct pca953x_chip *chip)
 {
 }
-#endif
 
-/*
- * Handlers for alternative sources of platform_data
- */
+#endif /* CONFIG_GPIO_PCA953X_IRQ */
+
 #ifdef CONFIG_OF_GPIO
+
 /*
  * Translate OpenFirmware node properties into platform_data
  */
 static struct pca953x_platform_data *
-pca953x_get_alt_pdata(struct i2c_client *client)
+pca953x_get_alt_pdata(struct pca953x_chip *chip)
 {
+	struct i2c_client *client = chip->client;
 	struct pca953x_platform_data *pdata;
 	struct device_node *node;
-	const uint16_t *val;
+	const int *val;
 
 	node = client->dev.of_node;
 	if (node == NULL)
@@ -469,19 +585,34 @@ pca953x_get_alt_pdata(struct i2c_client *client)
 			pdata->gpio_base = *val;
 	}
 
+	pdata->irq_base = -1;
+	val = of_get_property(node, "linux,irq-base", NULL);
+	if (val) {
+		if (*val < 0)
+			dev_warn(&client->dev,
+				 "invalid irq-base in device tree\n");
+		else
+			pdata->irq_base = *val;
+	}
+
 	val = of_get_property(node, "polarity", NULL);
 	if (val)
 		pdata->invert = *val;
 
+	pca953x_irq_alloc_host(chip, node);
+
 	return pdata;
 }
-#else
+
+#else /* CONFIG_OF_GPIO */
+
 static struct pca953x_platform_data *
 pca953x_get_alt_pdata(struct i2c_client *client)
 {
 	return NULL;
 }
-#endif
+
+#endif /* CONFIG_OF_GPIO */
 
 static int __devinit pca953x_probe(struct i2c_client *client,
 				   const struct i2c_device_id *id)
@@ -490,13 +621,18 @@ static int __devinit pca953x_probe(struct i2c_client *client,
 	struct pca953x_chip *chip;
 	int ret;
 
+	pr_debug("%s\n", __func__);
+
 	chip = kzalloc(sizeof(struct pca953x_chip), GFP_KERNEL);
 	if (chip == NULL)
 		return -ENOMEM;
 
+	chip->client = client;
+
 	pdata = client->dev.platform_data;
 	if (pdata == NULL) {
-		pdata = pca953x_get_alt_pdata(client);
+		pdata = pca953x_get_alt_pdata(chip);
+		client->dev.platform_data = pdata;
 		/*
 		 * Unlike normal platform_data, this is allocated
 		 * dynamically and must be freed in the driver
@@ -510,10 +646,7 @@ static int __devinit pca953x_probe(struct i2c_client *client,
 		goto out_failed;
 	}
 
-	chip->client = client;
-
 	chip->gpio_start = pdata->gpio_base;
-
 	chip->names = pdata->names;
 
 	/* initialize cached registers from their original values.
-- 
1.7.1

Re: [PATCH 2/2] gpio: pca953x: add powerpc irq support

[Thomas Gleixner]

On Fri, 4 Jun 2010, Esben Haabendal wrote:

NAK for various reasons (no particular order):

1) That patch misses a sensible changelog. See Documentation/Submit*

2) patch contains several independent changes, which need to be separated

> @@ -120,6 +124,10 @@ static int pca953x_gpio_direction_input(struct gpio_chip *gc, unsigned off)
>  	chip = container_of(gc, struct pca953x_chip, gpio_chip);
>  
>  	reg_val = chip->reg_direction | (1u << off);
> +
> +	if (reg_val == chip->reg_direction)
> +		return 0;
> +

 This is an optimization of its own value.

3) it breaks the driver. See http://lkml.org/lkml/2010/6/6/177 for a
   detailed explanation

4) the virq/powerpc churn is horrible and I bet there are sane ways to
   solve this, but it leave this to the powerpc experts.

Thanks,

	tglx

Re: [PATCH 2/2] gpio: pca953x: add powerpc irq support

[Esben Haabendal]

On Mon, 2010-06-07 at 01:39 +0200, Thomas Gleixner wrote: 
> On Fri, 4 Jun 2010, Esben Haabendal wrote:
> 
> NAK for various reasons (no particular order):

Ok. I can address 1 and 2, and I believe 3 is still under discussion, at
least as far as I am concerned.

> 1) That patch misses a sensible changelog. See Documentation/Submit*

I will address that.

> 2) patch contains several independent changes, which need to be separated

I will send a split up patch series. That is, if there is any chance
that it will be ACK'ed.  No need to waste time on it if it is dead
NAK'ed before arriving to list.

> > @@ -120,6 +124,10 @@ static int pca953x_gpio_direction_input(struct gpio_chip *gc, unsigned off)
> >  	chip = container_of(gc, struct pca953x_chip, gpio_chip);
> >  
> >  	reg_val = chip->reg_direction | (1u << off);
> > +
> > +	if (reg_val == chip->reg_direction)
> > +		return 0;
> > +
> 
>  This is an optimization of its own value.

Yes, but is need to avoid doing I2C work from irq_chip map().

> 3) it breaks the driver. See http://lkml.org/lkml/2010/6/6/177 for a
>    detailed explanation

I believe there is still a few things that needs to be discussed before
that is closed.

> 4) the virq/powerpc churn is horrible and I bet there are sane ways to
>    solve this, but it leave this to the powerpc experts.

Do you suggest that a seperate pca953x driver should be implemented for
powerpc?  (I guess not).  Or do you say that the who irq handling in
powerpc should be changed?

There must be an acceptable way to extend pca953x.c for the powerpc virq
handling and get it accepted in the kernel.

/Esben

Re: [PATCH 2/2] gpio: pca953x: add powerpc irq support

[Thomas Gleixner]

On Mon, 7 Jun 2010, Esben Haabendal wrote:
> On Mon, 2010-06-07 at 01:39 +0200, Thomas Gleixner wrote: 
> > On Fri, 4 Jun 2010, Esben Haabendal wrote:
>
> > > @@ -120,6 +124,10 @@ static int pca953x_gpio_direction_input(struct gpio_chip *gc, unsigned off)
> > >  	chip = container_of(gc, struct pca953x_chip, gpio_chip);
> > >  
> > >  	reg_val = chip->reg_direction | (1u << off);
> > > +
> > > +	if (reg_val == chip->reg_direction)
> > > +		return 0;
> > > +
> > 
> >  This is an optimization of its own value.
> 
> Yes, but is need to avoid doing I2C work from irq_chip map().

Still it should be a separate patch, that's all I said. 

It can be documented that it is also necessary to support powerpcs
virq thing if at all.

> > 3) it breaks the driver. See http://lkml.org/lkml/2010/6/6/177 for a
> >    detailed explanation
> 
> I believe there is still a few things that needs to be discussed before
> that is closed.

Not with me. Period.

> > 4) the virq/powerpc churn is horrible and I bet there are sane ways to
> >    solve this, but it leave this to the powerpc experts.
> 
> Do you suggest that a seperate pca953x driver should be implemented for
> powerpc?  (I guess not).  Or do you say that the who irq handling in
> powerpc should be changed?
> 
> There must be an acceptable way to extend pca953x.c for the powerpc virq
> handling and get it accepted in the kernel.

Yes, there certainly is. I just think that there are cleaner ways,
that's why I defered that to the powerpc wizards.

Thanks,

	tglx