mfd: twl: Fix for irqdomain/next + SPARSE_IRQ + MMC card detect

From: Felipe Balbi <[email protected]>

With sparse IRQs the driver shouldn't depend at all on any IRQ values coming from board-file.

Remove every occurences of pdata->irq_base/end.

From 70a1833ca20142ef5b6b96dbc92a63c7f8ef9a04 Mon Sep 17 00:00:00 2001

From d2b7b518b4c66beb99fccfcb898001c11a2f6dc9 Mon Sep 17 00:00:00 2001

Quoted 6 lines · Benoit Cousson

From: Felipe Balbi<[email protected]>

With sparse IRQs the driver shouldn't depend at all on any IRQ values coming from board-file.

Remove every occurences of pdata->irq_base/end.

Well, not quite *every*. :)

If the driver isn't going to use those fields anymore, they should also be removed from the pdata struct too[1]. Also, the remaining board/init code that is initializing those fields should be removed as well[2], since it obviously has no effect.

The first patch below should probably be folded into this one, and then the 2nd patch removes all the users.

Kevin

[1]
From 70a1833ca20142ef5b6b96dbc92a63c7f8ef9a04 Mon Sep 17 00:00:00 2001
From: Kevin Hilman<[email protected]>
Date: Wed, 14 Mar 2012 13:55:31 -0700
Subject: [PATCH 1/2] mfd: twl-core: remove pdata->irq_base/end

Signed-off-by Kevin Hilman<[email protected]>

---
  include/linux/i2c/twl.h |    1 -
  1 file changed, 1 deletion(-)

include/linux/i2c/twl.h +0 -1

index 2463b61..fa76006 100644
@@ -697,7 +697,6 @@ struct twl4030_audio_data {
  };
 
  struct twl4030_platform_data {
-	unsigned				irq_base, irq_end;
  	struct twl4030_clock_init_data		*clock;
  	struct twl4030_bci_platform_data	*bci;
  	struct twl4030_gpio_platform_data	*gpio;

Quoted 10 lines · Kevin Hilman

Quoted 6 lines · Benoit Cousson

From: Felipe Balbi<[email protected]>

With sparse IRQs the driver shouldn't depend at all on any IRQ values coming from board-file.

Remove every occurences of pdata->irq_base/end.

Well, not quite *every*. :)

Hehe, good catch, thanks.

Quoted 5 lines

If the driver isn't going to use those fields anymore, they should also be removed from the pdata struct too[1]. Also, the remaining board/init code that is initializing those fields should be removed as well[2], since it obviously has no effect.

Quoted 2 lines

The first patch below should probably be folded into this one, and then the 2nd patch removes all the users.

We should probably remove all the users before removing the attribute? It looks like your patches are ordered in the opposite way.

I'll add that in the series and repost the pull request.

Quoted 6 lines · Benoit Cousson

From: Felipe Balbi <[email protected]>

With sparse IRQs the driver shouldn't depend at all on any IRQ values coming from board-file.

Remove every occurences of pdata->irq_base/end.

Well, not quite *every*. :)

If the driver isn't going to use those fields anymore, they should also be removed from the pdata struct too[1]. Also, the remaining board/init code that is initializing those fields should be removed as well[2], since it obviously has no effect.

The first patch below should probably be folded into this one, and then the 2nd patch removes all the users.

From b59a4754fb66f6f729ebbd0563c4cd2bd2cda56e Mon Sep 17 00:00:00 2001

Quoted 11 lines · Kevin Hilman

Quoted 6 lines · Benoit Cousson

From: Felipe Balbi<[email protected]>

With sparse IRQs the driver shouldn't depend at all on any IRQ values coming from board-file.

Remove every occurences of pdata->irq_base/end.

Well, not quite *every*. :)

Quoted 7 lines

If the driver isn't going to use those fields anymore, they should also be removed from the pdata struct too[1]. Also, the remaining board/init code that is initializing those fields should be removed as well[2], since it obviously has no effect.

The first patch below should probably be folded into this one, and then the 2nd patch removes all the users.

Below is a correct version of the 2nd patch which removes all the users.

The previous version wouldn't apply cleanly because I generated it on top of a local branch that also included some other changes to twl-common.c from the SMPS regulator series (which seems to have missed the window for v3.4.)

This one applies cleanly on your for_3.4/twl_irq_gpio_mmc_fix branch.

Kevin

From b59a4754fb66f6f729ebbd0563c4cd2bd2cda56e Mon Sep 17 00:00:00 2001
From: Kevin Hilman<[email protected]>
Date: Wed, 14 Mar 2012 13:56:01 -0700
Subject: [PATCH] ARM: OMAP2+: TWL: remove usage of pdata->irq_base/_end

The driver has been converted to use SPARSE_IRQ and no longer needs to be passed IRQ base/end. The driver no longer uses these fields, so remove them from the reamaining users.

Signed-off-by Kevin Hilman<[email protected]>

---
  arch/arm/mach-omap2/board-2430sdp.c    |    3 ---
  arch/arm/mach-omap2/board-omap3logic.c |    3 ---
  arch/arm/mach-omap2/twl-common.c       |   10 ----------
  3 files changed, 16 deletions(-)

arch/arm/mach-omap2/board-2430sdp.c +0 -3

index 7370983..7dfcb96 100644
@@ -218,9 +218,6 @@ static struct twl4030_gpio_platform_data sdp2430_gpio_data = {
  };
 
  static struct twl4030_platform_data sdp2430_twldata = {
-	.irq_base	= TWL4030_IRQ_BASE,
-	.irq_end	= TWL4030_IRQ_END,
-
  	/* platform_data for children goes here */
  	.gpio		=&sdp2430_gpio_data,
  	.vmmc1		=&sdp2430_vmmc1,

arch/arm/mach-omap2/board-omap3logic.c +0 -3

index 4198dd0..0b053b4 100644
@@ -85,9 +85,6 @@ static struct twl4030_gpio_platform_data omap3logic_gpio_data = {
  };
 
  static struct twl4030_platform_data omap3logic_twldata = {
-	.irq_base	= TWL4030_IRQ_BASE,
-	.irq_end	= TWL4030_IRQ_END,
-
  	/* platform_data for children goes here */
  	.gpio		=&omap3logic_gpio_data,
  	.vmmc1		=&omap3logic_vmmc1,

arch/arm/mach-omap2/twl-common.c +0 -10

index 10b20c6..ce349ac 100644
@@ -129,11 +129,6 @@ static struct regulator_init_data omap3_vpll2_idata = {
  void __init omap3_pmic_get_config(struct twl4030_platform_data *pmic_data,
  				  u32 pdata_flags, u32 regulators_flags)
  {
-	if (!pmic_data->irq_base)
-		pmic_data->irq_base = TWL4030_IRQ_BASE;
-	if (!pmic_data->irq_end)
-		pmic_data->irq_end = TWL4030_IRQ_END;
-
  	/* Common platform data configurations */
  	if (pdata_flags&  TWL_COMMON_PDATA_USB&&  !pmic_data->usb)
  		pmic_data->usb =&omap3_usb_pdata;

@@ -287,11 +282,6 @@ static struct regulator_init_data omap4_clk32kg_idata = {
  void __init omap4_pmic_get_config(struct twl4030_platform_data *pmic_data,
  				  u32 pdata_flags, u32 regulators_flags)
  {
-	if (!pmic_data->irq_base)
-		pmic_data->irq_base = TWL6030_IRQ_BASE;
-	if (!pmic_data->irq_end)
-		pmic_data->irq_end = TWL6030_IRQ_END;
-
  	/* Common platform data configurations */
  	if (pdata_flags&  TWL_COMMON_PDATA_USB&&  !pmic_data->usb)
  		pmic_data->usb =&omap4_usb_pdata;

From: Kevin Hilman <[email protected]>
Date: Wed, 14 Mar 2012 13:56:01 -0700
Subject: [PATCH 2/2] ARM: OMAP2+: TWL: remove usage of pdata->irq_base/_end

The driver has been converted to use SPARSE_IRQ and no longer needs to be passed IRQ base/end. The driver no longer uses these fields, so remove them from the reamaining users.

Signed-off-by Kevin Hilman <[email protected]>

---
 arch/arm/mach-omap2/board-2430sdp.c    |    3 ---
 arch/arm/mach-omap2/board-omap3logic.c |    3 ---
 arch/arm/mach-omap2/twl-common.c       |    9 ---------
 3 files changed, 15 deletions(-)

arch/arm/mach-omap2/board-2430sdp.c +0 -3

index c8bda62..b8b3700 100644
@@ -218,9 +218,6 @@ static struct twl4030_gpio_platform_data sdp2430_gpio_data = {
 };
 
 static struct twl4030_platform_data sdp2430_twldata = {
-	.irq_base	= TWL4030_IRQ_BASE,
-	.irq_end	= TWL4030_IRQ_END,
-
 	/* platform_data for children goes here */
 	.gpio		= &sdp2430_gpio_data,
 	.vmmc1		= &sdp2430_vmmc1,

arch/arm/mach-omap2/board-omap3logic.c +0 -3

index 4a7d8c8..f145935 100644
@@ -85,9 +85,6 @@ static struct twl4030_gpio_platform_data omap3logic_gpio_data = {
 };
 
 static struct twl4030_platform_data omap3logic_twldata = {
-	.irq_base	= TWL4030_IRQ_BASE,
-	.irq_end	= TWL4030_IRQ_END,
-
 	/* platform_data for children goes here */
 	.gpio		= &omap3logic_gpio_data,
 	.vmmc1		= &omap3logic_vmmc1,

arch/arm/mach-omap2/twl-common.c +0 -4

index 1cddcb2..5c9aa3f 100644
@@ -184,10 +184,6 @@ static struct twl_regulator_driver_data omap3_vdd2_drvdata = {
 void __init omap3_pmic_get_config(struct twl4030_platform_data *pmic_data,
 				  u32 pdata_flags, u32 regulators_flags)
 {
-	if (!pmic_data->irq_base)
-		pmic_data->irq_base = TWL4030_IRQ_BASE;
-	if (!pmic_data->irq_end)
-		pmic_data->irq_end = TWL4030_IRQ_END;
 	if (!pmic_data->vdd1) {
 		omap3_vdd1.driver_data = &omap3_vdd1_drvdata;
 		omap3_vdd1_drvdata.data = voltdm_lookup("mpu_iva");

@@ -415,11 +411,6 @@ static struct twl_regulator_driver_data omap4_vdd3_drvdata = {
 void __init omap4_pmic_get_config(struct twl4030_platform_data *pmic_data,

Quoted 5 lines · Benoit Cousson

From: Felipe Balbi <[email protected]>

With sparse IRQs the driver shouldn't depend at all on any IRQ values coming from board-file.

Quoted 1 line

Remove every occurences of pdata->irq_base/end.

Well, not quite *every*. :)

If the driver isn't going to use those fields anymore, they should also be removed from the pdata struct too[1]. Also, the remaining board/init code that is initializing those fields should be removed as well[2], since it obviously has no effect.

Since a structure device is available now, use the dev_ macros instead of the pr_ ones.

Clean some badly formatted comments.
Remove some unused variables.
Move some variable to the place they belong.
Clean some badly wrapped lines.
Align variable definition
Add missing braces in if-then-else block.
Add blank line for better readability.
Move stuff here and there...

Signed-off-by Benoit Cousson<[email protected]>

Cc Felipe Balbi<[email protected]>

---
  drivers/mfd/twl-core.c    |   32 ++++++++++++++++++--------------
  drivers/mfd/twl4030-irq.c |   23 +++++++++++------------
  drivers/mfd/twl6030-irq.c |   39 +++++++++++++++++++--------------------
  3 files changed, 48 insertions(+), 46 deletions(-)

drivers/mfd/twl-core.c +18 -14

index 1c1b7d1..9bf14ef 100644
@@ -1174,17 +1174,15 @@ static int twl_remove(struct i2c_client *client)
  	return 0;
  }
 
-/* NOTE:  this driver only handles a single twl4030/tps659x0 chip */
+/* NOTE: This driver only handles a single twl4030/tps659x0 chip */
  static int __devinit
  twl_probe(struct i2c_client *client, const struct i2c_device_id *id)
  {
-	int				irq_base;
-	int				status;
-	unsigned			i;
  	struct twl4030_platform_data	*pdata = client->dev.platform_data;
  	struct device_node		*node = client->dev.of_node;
-	u8 temp;
-	int ret = 0;
+	int				irq_base = 0;
+	int				status;
+	unsigned			i;
 
  	if (node&&  !pdata) {
  		/*

@@ -1214,12 +1212,12 @@ twl_probe(struct i2c_client *client, const struct i2c_device_id *id)
  	}
 
  	for (i = 0; i<  TWL_NUM_SLAVES; i++) {
-		struct twl_client	*twl =&twl_modules[i];
+		struct twl_client *twl =&twl_modules[i];
 
  		twl->address = client->addr + i;
-		if (i == 0)
+		if (i == 0) {
  			twl->client = client;
-		else {
+		} else {
  			twl->client = i2c_new_dummy(client->adapter,
  					twl->address);
  			if (!twl->client) {

@@ -1231,7 +1229,9 @@ twl_probe(struct i2c_client *client, const struct i2c_device_id *id)
  		}
  		mutex_init(&twl->xfer_lock);
  	}
+
  	inuse = true;
+
  	if ((id->driver_data)&  TWL6030_CLASS) {
  		twl_id = TWL6030_CLASS_ID;
  		twl_map =&twl6030_map[0];

@@ -1245,8 +1245,8 @@ twl_probe(struct i2c_client *client, const struct i2c_device_id *id)
 
  	/* read TWL IDCODE Register */
  	if (twl_id == TWL4030_CLASS_ID) {
-		ret = twl_read_idcode_register();
-		WARN(ret<  0, "Error: reading twl_idcode register value\n");
+		status = twl_read_idcode_register();
+		WARN(status<  0, "Error: reading twl_idcode register value\n");
  	}
 
  	/* load power event scripts */

@@ -1268,19 +1268,22 @@ twl_probe(struct i2c_client *client, const struct i2c_device_id *id)
  		}
  	}
 
-	/* Disable TWL4030/TWL5030 I2C Pull-up on I2C1 and I2C4(SR) interface.
+	/*
+	 * Disable TWL4030/TWL5030 I2C Pull-up on I2C1 and I2C4(SR) interface.
  	 * Program I2C_SCL_CTRL_PU(bit 0)=0, I2C_SDA_CTRL_PU (bit 2)=0,
  	 * SR_I2C_SCL_CTRL_PU(bit 4)=0 and SR_I2C_SDA_CTRL_PU(bit 6)=0.
  	 */
-
  	if (twl_class_is_4030()) {
+		u8 temp;
+
  		twl_i2c_read_u8(TWL4030_MODULE_INTBR,&temp, REG_GPPUPDCTR1);
  		temp&= ~(SR_I2C_SDA_CTRL_PU | SR_I2C_SCL_CTRL_PU | \
-		I2C_SDA_CTRL_PU | I2C_SCL_CTRL_PU);
+			I2C_SDA_CTRL_PU | I2C_SCL_CTRL_PU);
  		twl_i2c_write_u8(TWL4030_MODULE_INTBR, temp, REG_GPPUPDCTR1);
  	}
 
  	status = -ENODEV;
+
  	if (node)
  		status = of_platform_populate(node, NULL, NULL,&client->dev);
  	if (status)

@@ -1289,6 +1292,7 @@ twl_probe(struct i2c_client *client, const struct i2c_device_id *id)
  fail:
  	if (status<  0)
  		twl_remove(client);
+
  	return status;
  }
 

drivers/mfd/twl4030-irq.c +11 -12

index e6dc8f9..e523499 100644
@@ -32,7 +32,6 @@
  #include<linux/interrupt.h>
  #include<linux/irq.h>
  #include<linux/slab.h>
-
  #include<linux/of.h>
  #include<linux/irqdomain.h>
  #include<linux/i2c/twl.h>

@@ -625,7 +624,8 @@ static irqreturn_t handle_twl4030_sih(int irq, void *data)
  	return IRQ_HANDLED;
  }
 
-/* returns the first IRQ used by this SIH bank,
+/*
+ * returns the first IRQ used by this SIH bank,
   * or negative errno
   */
  int twl4030_sih_setup(struct device *dev, int module, int irq_base)

@@ -637,14 +637,14 @@ int twl4030_sih_setup(struct device *dev, int module, int irq_base)
  	int			status = -EINVAL;
 
  	/* only support modules with standard clear-on-read for now */
-	for (sih_mod = 0, sih = sih_modules;
-			sih_mod<  nr_sih_modules;
+	for (sih_mod = 0, sih = sih_modules; sih_mod<  nr_sih_modules;
  			sih_mod++, sih++) {
  		if (sih->module == module&&  sih->set_cor) {
  			status = 0;
  			break;
  		}
  	}
+
  	if (status<  0)
  		return status;
 

@@ -674,7 +674,7 @@ int twl4030_sih_setup(struct device *dev, int module, int irq_base)
  	status = request_threaded_irq(irq, NULL, handle_twl4030_sih, 0,
  				      agent->irq_name ?: sih->name, NULL);
 
-	pr_info("twl4030: %s (irq %d) chaining IRQs %d..%d\n", sih->name,
+	dev_info(dev, "%s (irq %d) chaining IRQs %d..%d\n", sih->name,
  			irq, irq_base, irq_base + i - 1);
 
  	return status<  0 ? status : irq_base;

@@ -691,12 +691,10 @@ int twl4030_sih_setup(struct device *dev, int module, int irq_base)
  int twl4030_init_irq(struct device *dev, int irq_num)
  {
  	static struct irq_chip	twl4030_irq_chip;
+	int			status, i;
  	int			irq_base, irq_end, nr_irqs;
  	struct			device_node *node = dev->of_node;
 
-	int			status;
-	int			i;
-
  	/*
  	 * TWL core and pwr interrupts must be contiguous because
  	 * the hwirqs numbers are defined contiguously from 1 to 15.

@@ -725,7 +723,8 @@ int twl4030_init_irq(struct device *dev, int irq_num)
 
  	twl4030_irq_base = irq_base;
 
-	/* install an irq handler for each of the SIH modules;
+	/*
+	 * Install an irq handler for each of the SIH modules;
  	 * clone dummy irq_chip since PIH can't *do* anything
  	 */
  	twl4030_irq_chip = dummy_irq_chip;

@@ -740,13 +739,13 @@ int twl4030_init_irq(struct device *dev, int irq_num)
  		activate_irq(i);
  	}
 
-	pr_info("twl4030: %s (irq %d) chaining IRQs %d..%d\n", "PIH",
+	dev_info(dev, "%s (irq %d) chaining IRQs %d..%d\n", "PIH",
  			irq_num, irq_base, irq_end);
 
  	/* ... and the PWR_INT module ... */
  	status = twl4030_sih_setup(dev, TWL4030_MODULE_INT, irq_end);
  	if (status<  0) {
-		pr_err("twl4030: sih_setup PWR INT -->  %d\n", status);
+		dev_err(dev, "sih_setup PWR INT -->  %d\n", status);
  		goto fail;
  	}
 

@@ -755,7 +754,7 @@ int twl4030_init_irq(struct device *dev, int irq_num)
  				      IRQF_ONESHOT,
  				      "TWL4030-PIH", NULL);
  	if (status<  0) {
-		pr_err("twl4030: could not claim irq%d: %d\n", irq_num, status);
+		dev_err(dev, "could not claim irq%d: %d\n", irq_num, status);
  		goto fail_rqirq;
  	}
 

drivers/mfd/twl6030-irq.c +19 -20

index b4d5e0e..a1c7183 100644
@@ -53,7 +53,6 @@
   *
   * We set up IRQs starting at a platform-specified base. An interrupt map table,
   * specifies mapping between interrupt number and the associated module.
- *
   */
  #define TWL6030_NR_IRQS    20
 

@@ -342,14 +341,11 @@ int twl6030_init_irq(struct device *dev, int irq_num)
  {
  	struct			device_node *node = dev->of_node;
  	int			nr_irqs, irq_base, irq_end;
-
-	int	status = 0;
-	int	i;
  	struct task_struct	*task;
-	int ret;
-	u8 mask[4];
-
-	static struct irq_chip	twl6030_irq_chip;
+	static struct irq_chip  twl6030_irq_chip;
+	int			status = 0;
+	int			i;
+	u8			mask[4];
 
  	nr_irqs = TWL6030_NR_IRQS;
 

@@ -367,16 +363,18 @@ int twl6030_init_irq(struct device *dev, int irq_num)
  	mask[1] = 0xFF;
  	mask[2] = 0xFF;
  	mask[3] = 0xFF;
-	ret = twl_i2c_write(TWL_MODULE_PIH,&mask[0],
-			REG_INT_MSK_LINE_A, 3); /* MASK ALL INT LINES */
-	ret = twl_i2c_write(TWL_MODULE_PIH,&mask[0],
-			REG_INT_MSK_STS_A, 3); /* MASK ALL INT STS */
-	ret = twl_i2c_write(TWL_MODULE_PIH,&mask[0],
-			REG_INT_STS_A, 3); /* clear INT_STS_A,B,C */
+
+	/* mask all int lines */
+	twl_i2c_write(TWL_MODULE_PIH,&mask[0], REG_INT_MSK_LINE_A, 3);
+	/* mask all int sts */
+	twl_i2c_write(TWL_MODULE_PIH,&mask[0], REG_INT_MSK_STS_A, 3);
+	/* clear INT_STS_A,B,C */
+	twl_i2c_write(TWL_MODULE_PIH,&mask[0], REG_INT_STS_A, 3);
 
  	twl6030_irq_base = irq_base;
 
-	/* install an irq handler for each of the modules;
+	/*
+	 * install an irq handler for each of the modules;
  	 * clone dummy irq_chip since PIH can't *do* anything
  	 */
  	twl6030_irq_chip = dummy_irq_chip;

@@ -391,22 +389,22 @@ int twl6030_init_irq(struct device *dev, int irq_num)
  		activate_irq(i);
  	}
 
-	pr_info("twl6030: %s (irq %d) chaining IRQs %d..%d\n", "PIH",
+	dev_info(dev, "PIH (irq %d) chaining IRQs %d..%d\n",
  			irq_num, irq_base, irq_end);
 
  	/* install an irq handler to demultiplex the TWL6030 interrupt */
  	init_completion(&irq_event);
 
-	status = request_irq(irq_num, handle_twl6030_pih, 0,
-				"TWL6030-PIH",&irq_event);
+	status = request_irq(irq_num, handle_twl6030_pih, 0, "TWL6030-PIH",
+			&irq_event);
  	if (status<  0) {
-		pr_err("twl6030: could not claim irq%d: %d\n", irq_num, status);
+		dev_err(dev, "could not claim irq %d: %d\n", irq_num, status);
  		goto fail_irq;
  	}
 
  	task = kthread_run(twl6030_irq_thread, (void *)irq_num, "twl6030-irq");
  	if (IS_ERR(task)) {
-		pr_err("twl6030: could not create irq %d thread!\n", irq_num);
+		dev_err(dev, "could not create irq %d thread!\n", irq_num);
  		status = PTR_ERR(task);
  		goto fail_kthread;
  	}

@@ -421,6 +419,7 @@ fail_kthread:
  fail_irq:
  	for (i = irq_base; i<  irq_end; i++)
  		irq_set_chip_and_handler(i, NULL, NULL);
+
  	return status;
  }
 

Do not use the board pdata for irq_base, but allocate them dynamically to allow a proper support of SPARSE_IRQ.

Fix an unneeded line wrap.

Signed-off-by Benoit Cousson<[email protected]>

Acked-by Felipe Balbi<[email protected]>

---
  drivers/gpio/gpio-twl4030.c |   33 +++++++++++++++++----------------
  1 files changed, 17 insertions(+), 16 deletions(-)

drivers/gpio/gpio-twl4030.c +17 -16

index 697396c..49e5c6e 100644
@@ -395,23 +395,26 @@ static int gpio_twl4030_remove(struct platform_device *pdev);
  static int __devinit gpio_twl4030_probe(struct platform_device *pdev)
  {
  	struct twl4030_gpio_platform_data *pdata = pdev->dev.platform_data;
-	int ret;
+	int ret, irq_base;
 
  	/* maybe setup IRQs */
-	if (pdata->irq_base) {
-		if (is_module()) {
-			dev_err(&pdev->dev,
-				"can't dispatch IRQs from modules\n");
-			goto no_irqs;
-		}
-		ret = twl4030_sih_setup(&pdev->dev, TWL4030_MODULE_GPIO,
-					pdata->irq_base);
-		if (ret<  0)
-			return ret;
-		WARN_ON(ret != pdata->irq_base);
-		twl4030_gpio_irq_base = ret;
+	if (is_module()) {
+		dev_err(&pdev->dev, "can't dispatch IRQs from modules\n");
+		goto no_irqs;
+	}
+
+	irq_base = irq_alloc_descs(-1, 0, TWL4030_GPIO_MAX, 0);
+	if (irq_base<  0) {
+		dev_err(&pdev->dev, "Failed to alloc irq_descs\n");
+		return irq_base;
  	}
 
+	ret = twl4030_sih_setup(&pdev->dev, TWL4030_MODULE_GPIO, irq_base);
+	if (ret<  0)
+		return ret;
+
+	twl4030_gpio_irq_base = irq_base;
+
  no_irqs:
  	/*
  	 * NOTE:  boards may waste power if they don't set pullups

@@ -443,9 +446,7 @@ no_irqs:
 
  	ret = gpiochip_add(&twl_gpiochip);
  	if (ret<  0) {
-		dev_err(&pdev->dev,
-				"could not register gpiochip, %d\n",
-				ret);
+		dev_err(&pdev->dev, "could not register gpiochip, %d\n", ret);
  		twl_gpiochip.ngpio = 0;
  		gpio_twl4030_remove(pdev);
  	} else if (pdata->setup) {

Hi Grant,

That fix is tightly coupled with the previous twl4030-irq change, so it will be good to pulled it with the twl series through MFD if you are OK with that?

Care to ack this one and the next one?

Thanks,
Benoit

Quoted 57 lines · Benoit Cousson

Do not use the board pdata for irq_base, but allocate them dynamically to allow a proper support of SPARSE_IRQ.

Fix an unneeded line wrap.

Signed-off-by Benoit Cousson<[email protected]>

Acked-by Felipe Balbi<[email protected]>

---
drivers/gpio/gpio-twl4030.c | 33 +++++++++++++++++----------------

1 files changed, 17 insertions(+), 16 deletions(-)

drivers/gpio/gpio-twl4030.c +16 -13

index 697396c..49e5c6e 100644
@@ -395,23 +395,26 @@ static int gpio_twl4030_remove(struct
 platform_device *pdev);
 static int __devinit gpio_twl4030_probe(struct platform_device *pdev)
 {
 struct twl4030_gpio_platform_data *pdata = pdev->dev.platform_data;
- int ret;
+ int ret, irq_base;
 
 /* maybe setup IRQs */
- if (pdata->irq_base) {
- if (is_module()) {
- dev_err(&pdev->dev,
- "can't dispatch IRQs from modules\n");
- goto no_irqs;
- }
- ret = twl4030_sih_setup(&pdev->dev, TWL4030_MODULE_GPIO,
- pdata->irq_base);
- if (ret< 0)
- return ret;
- WARN_ON(ret != pdata->irq_base);
- twl4030_gpio_irq_base = ret;
+ if (is_module()) {
+ dev_err(&pdev->dev, "can't dispatch IRQs from modules\n");
+ goto no_irqs;
+ }
+
+ irq_base = irq_alloc_descs(-1, 0, TWL4030_GPIO_MAX, 0);
+ if (irq_base< 0) {
+ dev_err(&pdev->dev, "Failed to alloc irq_descs\n");
+ return irq_base;
 }
 
+ ret = twl4030_sih_setup(&pdev->dev, TWL4030_MODULE_GPIO, irq_base);
+ if (ret< 0)
+ return ret;
+
+ twl4030_gpio_irq_base = irq_base;
+

no_irqs:
/*
* NOTE: boards may waste power if they don't set pullups

drivers/gpio/gpio-twl4030.c no_irqs: · line 452

Quoted 10 lines

drivers/gpio/gpio-twl4030.c +1 -3

@@ -443,9 +446,7 @@ no_irqs:
 
 ret = gpiochip_add(&twl_gpiochip);
 if (ret< 0) {
- dev_err(&pdev->dev,
- "could not register gpiochip, %d\n",
- ret);
+ dev_err(&pdev->dev, "could not register gpiochip, %d\n", ret);
 twl_gpiochip.ngpio = 0;
 gpio_twl4030_remove(pdev);
 } else if (pdata->setup) {

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

Hi Grant,

That fix is tightly coupled with the previous twl4030-irq change, so it will be good to pulled it with the twl series through MFD if you are OK with that?

Care to ack this one and the next one?

Thanks,
Benoit

Quoted 7 lines · Benoit Cousson

Do not use the board pdata for irq_base, but allocate them dynamically to allow a proper support of SPARSE_IRQ.

Fix an unneeded line wrap.

Signed-off-by Benoit Cousson<[email protected]>

Acked-by Felipe Balbi<[email protected]>

Quoted 59 lines

---
  drivers/gpio/gpio-twl4030.c |   33 +++++++++++++++++----------------
  1 files changed, 17 insertions(+), 16 deletions(-)

drivers/gpio/gpio-twl4030.c +17 -16

index 697396c..49e5c6e 100644
@@ -395,23 +395,26 @@ static int gpio_twl4030_remove(struct platform_device *pdev);
  static int __devinit gpio_twl4030_probe(struct platform_device *pdev)
  {
  	struct twl4030_gpio_platform_data *pdata = pdev->dev.platform_data;
-	int ret;
+	int ret, irq_base;
 
  	/* maybe setup IRQs */
-	if (pdata->irq_base) {
-		if (is_module()) {
-			dev_err(&pdev->dev,
-				"can't dispatch IRQs from modules\n");
-			goto no_irqs;
-		}
-		ret = twl4030_sih_setup(&pdev->dev, TWL4030_MODULE_GPIO,
-					pdata->irq_base);
-		if (ret<  0)
-			return ret;
-		WARN_ON(ret != pdata->irq_base);
-		twl4030_gpio_irq_base = ret;
+	if (is_module()) {
+		dev_err(&pdev->dev, "can't dispatch IRQs from modules\n");
+		goto no_irqs;
+	}
+
+	irq_base = irq_alloc_descs(-1, 0, TWL4030_GPIO_MAX, 0);
+	if (irq_base<  0) {
+		dev_err(&pdev->dev, "Failed to alloc irq_descs\n");
+		return irq_base;
  	}
 
+	ret = twl4030_sih_setup(&pdev->dev, TWL4030_MODULE_GPIO, irq_base);
+	if (ret<  0)
+		return ret;
+
+	twl4030_gpio_irq_base = irq_base;
+
  no_irqs:
  	/*
  	 * NOTE:  boards may waste power if they don't set pullups

@@ -443,9 +446,7 @@ no_irqs:
 
  	ret = gpiochip_add(&twl_gpiochip);
  	if (ret<  0) {
-		dev_err(&pdev->dev,
-				"could not register gpiochip, %d\n",
-				ret);
+		dev_err(&pdev->dev, "could not register gpiochip, %d\n", ret);
  		twl_gpiochip.ngpio = 0;
  		gpio_twl4030_remove(pdev);
  	} else if (pdata->setup) {

Add the DT support for the I2C GPIO expander inside the twl4030.

Signed-off-by Benoit Cousson <[email protected]>

Acked-by Felipe Balbi <[email protected]>

---
 .../devicetree/bindings/gpio/gpio-twl4030.txt      |   23 ++++++
 drivers/gpio/gpio-twl4030.c                        |   79 ++++++++++++-------
 2 files changed, 73 insertions(+), 29 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/gpio/gpio-twl4030.txt

Documentation/devicetree/bindings/gpio/gpio-twl4030.txt new file +23 -0

new file mode 100644
index 0000000..16695d9
@@ -0,0 +1,23 @@
+twl4030 GPIO controller bindings
+
+Required properties:
+- compatible:
+  - "ti,twl4030-gpio" for twl4030 GPIO controller
+- #gpio-cells : Should be two.
+  - first cell is the pin number
+  - second cell is used to specify optional parameters (unused)
+- gpio-controller : Marks the device node as a GPIO controller.
+- #interrupt-cells : Should be 2.
+- interrupt-controller: Mark the device node as an interrupt controller
+  The first cell is the GPIO number.
+  The second cell is not used.
+
+Example:
+
+twl_gpio: gpio {
+    compatible = "ti,twl4030-gpio";
+    #gpio-cells = <2>;
+    gpio-controller;
+    #interrupt-cells = <2>;
+    interrupt-controller;
+};

drivers/gpio/gpio-twl4030.c +50 -29

index 49e5c6e..94256fe 100644
@@ -32,6 +32,8 @@
 #include <linux/irq.h>
 #include <linux/gpio.h>
 #include <linux/platform_device.h>
+#include <linux/of.h>
+#include <linux/irqdomain.h>
 
 #include <linux/i2c/twl.h>
 

@@ -256,7 +258,8 @@ static int twl_request(struct gpio_chip *chip, unsigned offset)
 		 * and vMMC2 power supplies based on card presence.
 		 */
 		pdata = chip->dev->platform_data;
-		value |= pdata->mmc_cd & 0x03;
+		if (pdata)
+			value |= pdata->mmc_cd & 0x03;
 
 		status = gpio_twl4030_write(REG_GPIO_CTRL, value);
 	}

@@ -395,6 +398,7 @@ static int gpio_twl4030_remove(struct platform_device *pdev);
 static int __devinit gpio_twl4030_probe(struct platform_device *pdev)
 {
 	struct twl4030_gpio_platform_data *pdata = pdev->dev.platform_data;
+	struct device_node *node = pdev->dev.of_node;
 	int ret, irq_base;
 
 	/* maybe setup IRQs */

@@ -409,6 +413,9 @@ static int __devinit gpio_twl4030_probe(struct platform_device *pdev)
 		return irq_base;
 	}
 
+	irq_domain_add_legacy(node, TWL4030_GPIO_MAX, irq_base, 0,
+			      &irq_domain_simple_ops, NULL);
+
 	ret = twl4030_sih_setup(&pdev->dev, TWL4030_MODULE_GPIO, irq_base);
 	if (ret < 0)
 		return ret;

@@ -416,40 +423,45 @@ static int __devinit gpio_twl4030_probe(struct platform_device *pdev)
 	twl4030_gpio_irq_base = irq_base;
 
 no_irqs:
-	/*
-	 * NOTE:  boards may waste power if they don't set pullups
-	 * and pulldowns correctly ... default for non-ULPI pins is
-	 * pulldown, and some other pins may have external pullups
-	 * or pulldowns.  Careful!
-	 */
-	ret = gpio_twl4030_pulls(pdata->pullups, pdata->pulldowns);
-	if (ret)
-		dev_dbg(&pdev->dev, "pullups %.05x %.05x --> %d\n",
-				pdata->pullups, pdata->pulldowns,
-				ret);
-
-	ret = gpio_twl4030_debounce(pdata->debounce, pdata->mmc_cd);
-	if (ret)
-		dev_dbg(&pdev->dev, "debounce %.03x %.01x --> %d\n",
-				pdata->debounce, pdata->mmc_cd,
-				ret);
-
-	twl_gpiochip.base = pdata->gpio_base;
+	twl_gpiochip.base = -1;
 	twl_gpiochip.ngpio = TWL4030_GPIO_MAX;
 	twl_gpiochip.dev = &pdev->dev;
 
-	/* NOTE: we assume VIBRA_CTL.VIBRA_EN, in MODULE_AUDIO_VOICE,
-	 * is (still) clear if use_leds is set.
-	 */
-	if (pdata->use_leds)
-		twl_gpiochip.ngpio += 2;
+	if (pdata) {
+		twl_gpiochip.base = pdata->gpio_base;
+
+		/*
+		 * NOTE:  boards may waste power if they don't set pullups
+		 * and pulldowns correctly ... default for non-ULPI pins is
+		 * pulldown, and some other pins may have external pullups
+		 * or pulldowns.  Careful!
+		 */
+		ret = gpio_twl4030_pulls(pdata->pullups, pdata->pulldowns);
+		if (ret)
+			dev_dbg(&pdev->dev, "pullups %.05x %.05x --> %d\n",
+					pdata->pullups, pdata->pulldowns,
+					ret);
+
+		ret = gpio_twl4030_debounce(pdata->debounce, pdata->mmc_cd);
+		if (ret)
+			dev_dbg(&pdev->dev, "debounce %.03x %.01x --> %d\n",
+					pdata->debounce, pdata->mmc_cd,
+					ret);
+
+		/*
+		 * NOTE: we assume VIBRA_CTL.VIBRA_EN, in MODULE_AUDIO_VOICE,
+		 * is (still) clear if use_leds is set.
+		 */
+		if (pdata->use_leds)
+			twl_gpiochip.ngpio += 2;
+	}
 
 	ret = gpiochip_add(&twl_gpiochip);
 	if (ret < 0) {
 		dev_err(&pdev->dev, "could not register gpiochip, %d\n", ret);
 		twl_gpiochip.ngpio = 0;
 		gpio_twl4030_remove(pdev);
-	} else if (pdata->setup) {
+	} else if (pdata && pdata->setup) {
 		int status;
 
 		status = pdata->setup(&pdev->dev,

@@ -467,7 +479,7 @@ static int gpio_twl4030_remove(struct platform_device *pdev)
 	struct twl4030_gpio_platform_data *pdata = pdev->dev.platform_data;
 	int status;
 
-	if (pdata->teardown) {
+	if (pdata && pdata->teardown) {
 		status = pdata->teardown(&pdev->dev,
 				pdata->gpio_base, TWL4030_GPIO_MAX);
 		if (status) {

@@ -488,12 +500,21 @@ static int gpio_twl4030_remove(struct platform_device *pdev)
 	return -EIO;
 }
 
+static const struct of_device_id twl_gpio_match[] = {
+	{ .compatible = "ti,twl4030-gpio", },
+	{ },
+};
+MODULE_DEVICE_TABLE(of, twl_gpio_match);
+
 /* Note:  this hardware lives inside an I2C-based multi-function device. */
 MODULE_ALIAS("platform:twl4030_gpio");
 
 static struct platform_driver gpio_twl4030_driver = {
-	.driver.name	= "twl4030_gpio",
-	.driver.owner	= THIS_MODULE,
+	.driver = {
+		.name	= "twl4030_gpio",
+		.owner	= THIS_MODULE,
+		.of_match_table = of_match_ptr(twl_gpio_match),
+	},
 	.probe		= gpio_twl4030_probe,
 	.remove		= gpio_twl4030_remove,
 };

Signature · 1 line

1.7.0.4

Card detect IRQ from the TWL6030 used to be provided to the MMC controller code using a statically allocated IRQ scheme:

  card_detect_irq = TWL6030_IRQ_BASE + MMCDETECT_INTR_OFFSET;

This is no longer valid in a SPARSE_IRQ context since there is no more pre-defined TWL6030_IRQ_BASE.

Return the proper card detect IRQ value in the twl6030_mmc_card_detect_config that will be called from the MMC controller.

Signed-off-by Benoit Cousson <[email protected]>

Cc Felipe Balbi <[email protected]>

Cc Tony Lindgren <[email protected]>

Cc Rajendra Nayak <[email protected]>

---
 drivers/mfd/twl6030-irq.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

drivers/mfd/twl6030-irq.c +0 -0

index a1c7183..4a63de2 100644

Hi,

Quoted 8 lines · Benoit Cousson

Card detect IRQ from the TWL6030 used to be provided to the MMC controller code using a statically allocated IRQ scheme:

  card_detect_irq = TWL6030_IRQ_BASE + MMCDETECT_INTR_OFFSET;

This is no longer valid in a SPARSE_IRQ context since there is no more pre-defined TWL6030_IRQ_BASE.

Quoted 7 lines

Return the proper card detect IRQ value in the twl6030_mmc_card_detect_config that will be called from the MMC controller.

Signed-off-by Benoit Cousson <[email protected]>

Cc Felipe Balbi <[email protected]>

Cc Tony Lindgren <[email protected]>

Cc Rajendra Nayak <[email protected]>

Reviewed-by Felipe Balbi <[email protected]>

Quoted 9 lines

---
 drivers/mfd/twl6030-irq.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

drivers/mfd/twl6030-irq.c +0 -0

index a1c7183..4a63de2 100644

drivers/mfd/twl6030-irq.c int twl6030_mmc_card_detect_config(void) · line 314

Quoted 8 lines

drivers/mfd/twl6030-irq.c +2 -1

@@ -308,7 +308,8 @@ int twl6030_mmc_card_detect_config(void)
 									ret);
 		return ret;
 	}
-	return 0;
+
+	return twl6030_irq_base + MMCDETECT_INTR_OFFSET;
 }
 EXPORT_SYMBOL(twl6030_mmc_card_detect_config);

At some point this twl6030_mmc_card_detect_config() needs to vanish in favor of a better mechanism. Maybe passing the GPIO number an requiring the driver to request the GPIO, set its direction and use it as IRQ line.

the only problem would arise if we end up hooking the Card Detect pin to a real IRQ line, but then we can use a flag to differentiate.

Quoted 44 lines · Felipe Balbi

Hi,

Quoted 8 lines · Benoit Cousson

Card detect IRQ from the TWL6030 used to be provided to the MMC controller code using a statically allocated IRQ scheme:

   card_detect_irq = TWL6030_IRQ_BASE + MMCDETECT_INTR_OFFSET;

This is no longer valid in a SPARSE_IRQ context since there is no more pre-defined TWL6030_IRQ_BASE.

Quoted 7 lines

Return the proper card detect IRQ value in the twl6030_mmc_card_detect_config that will be called from the MMC controller.

Signed-off-by Benoit Cousson<[email protected]>

Cc Felipe Balbi<[email protected]>

Cc Tony Lindgren<[email protected]>

Cc Rajendra Nayak<[email protected]>

Reviewed-by Felipe Balbi<[email protected]>

Quoted 9 lines

---
  drivers/mfd/twl6030-irq.c |    3 ++-
  1 files changed, 2 insertions(+), 1 deletions(-)

drivers/mfd/twl6030-irq.c +0 -0

index a1c7183..4a63de2 100644

drivers/mfd/twl6030-irq.c int twl6030_mmc_card_detect_config(void) · line 314

Quoted 8 lines

drivers/mfd/twl6030-irq.c +2 -1

@@ -308,7 +308,8 @@ int twl6030_mmc_card_detect_config(void)
  									ret);
  		return ret;
  	}
-	return 0;
+
+	return twl6030_irq_base + MMCDETECT_INTR_OFFSET;
  }
  EXPORT_SYMBOL(twl6030_mmc_card_detect_config);

At some point this twl6030_mmc_card_detect_config() needs to vanish in favor of a better mechanism. Maybe passing the GPIO number an requiring the driver to request the GPIO, set its direction and use it as IRQ line.

Quoted 2 lines

the only problem would arise if we end up hooking the Card Detect pin to a real IRQ line, but then we can use a flag to differentiate.

I'm trying to leave out the need for this callback to pdata.. See mmc: omap_hsmmc: Simplify init for twl6030 MMC card detect that does pretty much the same as Benoit's patch plus removes the need for the callback.

Hi Tony,

Quoted 46 lines · Tony Lindgren

Quoted 45 lines · Felipe Balbi

Hi,

Quoted 8 lines · Benoit Cousson

Card detect IRQ from the TWL6030 used to be provided to the MMC controller code using a statically allocated IRQ scheme:

card_detect_irq = TWL6030_IRQ_BASE + MMCDETECT_INTR_OFFSET;

This is no longer valid in a SPARSE_IRQ context since there is no more pre-defined TWL6030_IRQ_BASE.

Quoted 8 lines

Return the proper card detect IRQ value in the
twl6030_mmc_card_detect_config
that will be called from the MMC controller.

Signed-off-by Benoit Cousson<[email protected]>

Cc Felipe Balbi<[email protected]>

Cc Tony Lindgren<[email protected]>

Cc Rajendra Nayak<[email protected]>

Reviewed-by Felipe Balbi<[email protected]>

Quoted 9 lines

---
drivers/mfd/twl6030-irq.c | 3 ++-

1 files changed, 2 insertions(+), 1 deletions(-)

drivers/mfd/twl6030-irq.c +0 -0

index a1c7183..4a63de2 100644

drivers/mfd/twl6030-irq.c int twl6030_mmc_card_detect_config(void) · line 314

Quoted 8 lines

drivers/mfd/twl6030-irq.c +2 -1

@@ -308,7 +308,8 @@ int twl6030_mmc_card_detect_config(void)
 ret);
 return ret;
 }
- return 0;
+
+ return twl6030_irq_base + MMCDETECT_INTR_OFFSET;
 }
 EXPORT_SYMBOL(twl6030_mmc_card_detect_config);

At some point this twl6030_mmc_card_detect_config() needs to vanish in favor of a better mechanism. Maybe passing the GPIO number an requiring the driver to request the GPIO, set its direction and use it as IRQ line.

Quoted 7 lines

Quoted 2 lines

the only problem would arise if we end up hooking the Card Detect pin to a real IRQ line, but then we can use a flag to differentiate.

I'm trying to leave out the need for this callback to pdata.. See mmc: omap_hsmmc: Simplify init for twl6030 MMC card detect that does pretty much the same as Benoit's patch plus removes the need for the callback.

That's cool. I did that patch to still have a working MMC after my TWL series because I missed your MMC series :-(.

The patch #4 "mmc: omap_hsmmc: Simplify init for twl6030 MMC card detect" is indeed simplifying a lot that IRQ mess.

I will check with that TWL series, but since you are already using the twl6030_irq_base it should work without any trouble.

I created a for_3.4/twl_irq_gpio_fix branch based on irqdomain/next with all the TWL + GPIO patches except the DTS and without the MMC hacks.

And I merged it into your hsmmc-gpio branch. There is a small conflict in gpio-twl4030 but otherwise it works on the SDP4430.

I cannot test remotely with Panda or Beagle, the animals are already sleeping.

The branch is there if you want to check for OMAP3 as well.
git://git.kernel.org/pub/scm/linux/kernel/git/bcousson/linux-omap-dt.git
for_3.4/twl_irq_gpio_fix

Hi,

Quoted 8 lines · Benoit Cousson

Card detect IRQ from the TWL6030 used to be provided to the MMC controller code using a statically allocated IRQ scheme:

   card_detect_irq = TWL6030_IRQ_BASE + MMCDETECT_INTR_OFFSET;

This is no longer valid in a SPARSE_IRQ context since there is no more pre-defined TWL6030_IRQ_BASE.

Quoted 7 lines

Return the proper card detect IRQ value in the twl6030_mmc_card_detect_config that will be called from the MMC controller.

Signed-off-by Benoit Cousson<[email protected]>

Cc Felipe Balbi<[email protected]>

Cc Tony Lindgren<[email protected]>

Cc Rajendra Nayak<[email protected]>

Reviewed-by Felipe Balbi<[email protected]>

Quoted 9 lines

---
  drivers/mfd/twl6030-irq.c |    3 ++-
  1 files changed, 2 insertions(+), 1 deletions(-)

drivers/mfd/twl6030-irq.c +0 -0

index a1c7183..4a63de2 100644

drivers/mfd/twl6030-irq.c int twl6030_mmc_card_detect_config(void) · line 314

Quoted 8 lines

drivers/mfd/twl6030-irq.c +2 -1

@@ -308,7 +308,8 @@ int twl6030_mmc_card_detect_config(void)
  									ret);
  		return ret;
  	}
-	return 0;
+
+	return twl6030_irq_base + MMCDETECT_INTR_OFFSET;
  }
  EXPORT_SYMBOL(twl6030_mmc_card_detect_config);

At some point this twl6030_mmc_card_detect_config() needs to vanish in favor of a better mechanism. Maybe passing the GPIO number an requiring the driver to request the GPIO, set its direction and use it as IRQ line.

the only problem would arise if we end up hooking the Card Detect pin to a real IRQ line, but then we can use a flag to differentiate.

Quoted 45 lines · Felipe Balbi

Hi,

Quoted 8 lines · Benoit Cousson

Card detect IRQ from the TWL6030 used to be provided to the MMC controller code using a statically allocated IRQ scheme:

card_detect_irq = TWL6030_IRQ_BASE + MMCDETECT_INTR_OFFSET;

This is no longer valid in a SPARSE_IRQ context since there is no more pre-defined TWL6030_IRQ_BASE.

Quoted 8 lines

Return the proper card detect IRQ value in the
twl6030_mmc_card_detect_config
that will be called from the MMC controller.

Signed-off-by Benoit Cousson<[email protected]>

Cc Felipe Balbi<[email protected]>

Cc Tony Lindgren<[email protected]>

Cc Rajendra Nayak<[email protected]>

Reviewed-by Felipe Balbi<[email protected]>

Quoted 9 lines

---
drivers/mfd/twl6030-irq.c | 3 ++-

1 files changed, 2 insertions(+), 1 deletions(-)

drivers/mfd/twl6030-irq.c +0 -0

index a1c7183..4a63de2 100644

drivers/mfd/twl6030-irq.c int twl6030_mmc_card_detect_config(void) · line 314

Quoted 8 lines

drivers/mfd/twl6030-irq.c +2 -1

@@ -308,7 +308,8 @@ int twl6030_mmc_card_detect_config(void)
 ret);
 return ret;
 }
- return 0;
+
+ return twl6030_irq_base + MMCDETECT_INTR_OFFSET;
 }
 EXPORT_SYMBOL(twl6030_mmc_card_detect_config);

At some point this twl6030_mmc_card_detect_config() needs to vanish in favor of a better mechanism. Maybe passing the GPIO number an requiring the driver to request the GPIO, set its direction and use it as IRQ line.

Quoted 2 lines

the only problem would arise if we end up hooking the Card Detect pin to a real IRQ line, but then we can use a flag to differentiate.

Shouldn't this be hidden from the driver completely if the card detect pin is a GPIO or connected to a real IRQ line? And the board/DT handle this instead of the driver knowing when to do a gpio_to_irq()?

Card detect IRQ from the TWL6030 used to be provided to the MMC

controller code using a statically allocated IRQ scheme:

   card_detect_irq = TWL6030_IRQ_BASE + MMCDETECT_INTR_OFFSET;

This is no longer valid in a SPARSE_IRQ context since there is no more pre-defined TWL6030_IRQ_BASE.

Return the proper card detect IRQ value in the twl6030_mmc_card_detect_config that will be called from the MMC controller.

Signed-off-by Benoit Cousson<[email protected]>

Cc Felipe Balbi<[email protected]>

Cc Tony Lindgren<[email protected]>

Cc Rajendra Nayak<[email protected]>

---
  drivers/mfd/twl6030-irq.c |    3 ++-

Hello.

Quoted 13 lines · Benoit Cousson

Card detect IRQ from the TWL6030 used to be provided to the MMC

controller code using a statically allocated IRQ scheme:

card_detect_irq = TWL6030_IRQ_BASE + MMCDETECT_INTR_OFFSET;

This is no longer valid in a SPARSE_IRQ context since there is no more pre-defined TWL6030_IRQ_BASE.

Return the proper card detect IRQ value in the
twl6030_mmc_card_detect_config
that will be called from the MMC controller.

Quoted 6 lines

Signed-off-by Benoit Cousson<[email protected]>

Cc Felipe Balbi<[email protected]>

Cc Tony Lindgren<[email protected]>

Cc Rajendra Nayak<[email protected]>

---
drivers/mfd/twl6030-irq.c | 3 ++-

Subject talks about twl4030-irq.

With the introduction of dynamically allocated IRQ in the twl6030 driver, the board files can no longer rely of static IRQ defines like before.

Retrieve the value dynamically allocated from the mmc -> twl6030 init callback.

Note: The Panda board does not seems to use the card detect IRQ.

Quoted 6 lines · Benoit Cousson

With the introduction of dynamically allocated IRQ in the twl6030 driver, the board files can no longer rely of static IRQ defines like before.

Retrieve the value dynamically allocated from the mmc -> twl6030 init callback.

Quoted 1 line

Note: The Panda board does not seems to use the card detect IRQ.

Looks like this one is safe to merge via MFD and does not seem to cause any merge conflicts.

This one should be folded into patch 08/12 to avoid boot issues for between patches 08 and 12 in this series. Other than that:

Acked-by Tony Lindgren<[email protected]>

This is not for pull-request, becasue it is based on irqdomain + OMAP IRQ DT series + OMAP twl DT series yet to be pushed.

Hi Benoit,

Benoit Cousson<[email protected]> writes:

[...]

Quoted 2 lines

This is not for pull-request, becasue it is based on irqdomain + OMAP IRQ DT series + OMAP twl DT series yet to be pushed.

What's the status of this series?

At least 'gpio/twl: Allocate irq_desc dynamically for SPARSE_IRQ support' is needed in mainline (linux-next) now since without it, non-DT boots will dump a big fat warning trace[2].

Also FYI, I needed the patch below[1] in order to build this series.

Kevin

[1]

arch/arm/mach-omap2/board-generic.c +1 -0

index b8c74c7..8db1d64 100644
@@ -15,6 +15,7 @@
  #include<linux/of_irq.h>
  #include<linux/of_platform.h>
  #include<linux/irqdomain.h>
+#include<linux/i2c/twl.h>
 
  #include<mach/hardware.h>
  #include<asm/hardware/gic.h>

[2]

[    0.495117] twl4030: PIH (irq 7) chaining IRQs 368..401
[    0.501190] twl4030: power (irq 373) chaining IRQs 402..409
[    0.509429] twl4030: gpio (irq 368) chaining IRQs 410..427
[    0.515319] ------------[ cut here ]------------
[    0.520202] WARNING: at /work/kernel/omap/pm/drivers/gpio/gpio-twl4030.c:410 gpio_twl4030_probe+0x44/0x214()
[    0.530395] Modules linked in:
[    0.533691] [<c0014248>] (unwind_backtrace+0x0/0xf0) from [<c0039054>] (warn_slowpath_common+0x4c/0x64)
[    0.543640] [<c0039054>] (warn_slowpath_common+0x4c/0x64) from [<c0039088>] (warn_slowpath_null+0x1c/0x24)
[    0.553680] [<c0039088>] (warn_slowpath_null+0x1c/0x24) from [<c04196f4>] (gpio_twl4030_probe+0x44/0x214)
[    0.563629] [<c04196f4>] (gpio_twl4030_probe+0x44/0x214) from [<c028832c>] (platform_drv_probe+0x18/0x1c)
[    0.573577] [<c028832c>] (platform_drv_probe+0x18/0x1c) from [<c0286fd4>] (really_probe+0x60/0x15c)
[    0.583007] [<c0286fd4>] (really_probe+0x60/0x15c) from [<c0287214>] (driver_probe_device+0x48/0x60)
[    0.592498] [<c0287214>] (driver_probe_device+0x48/0x60) from [<c0285af8>] (bus_for_each_drv+0x5c/0x88)
[    0.602264] [<c0285af8>] (bus_for_each_drv+0x5c/0x88) from [<c028718c>] (device_attach+0x98/0xbc)
[    0.611480] [<c028718c>] (device_attach+0x98/0xbc) from [<c0286844>] (bus_probe_device+0x88/0xac)
[    0.620727] [<c0286844>] (bus_probe_device+0x88/0xac) from [<c0285160>] (device_add+0x278/0x358)
[    0.629882] [<c0285160>] (device_add+0x278/0x358) from [<c02887a8>] (platform_device_add+0xf8/0x1a4)
[    0.639373] [<c02887a8>] (platform_device_add+0xf8/0x1a4) from [<c042e4a0>] (add_numbered_child.constprop.0+0xb8/0xfc)
[    0.650451] [<c042e4a0>] (add_numbered_child.constprop.0+0xb8/0xfc) from [<c042e5a4>] (add_children+0x44/0x6d0)
[    0.660949] [<c042e5a4>] (add_children+0x44/0x6d0) from [<c041a8b8>] (twl_probe+0x354/0x3bc)
[    0.669738] [<c041a8b8>] (twl_probe+0x354/0x3bc) from [<c0326cb0>] (i2c_device_probe+0xc0/0x100)
[    0.678863] [<c0326cb0>] (i2c_device_probe+0xc0/0x100) from [<c0286fd4>] (really_probe+0x60/0x15c)
[    0.688171] [<c0286fd4>] (really_probe+0x60/0x15c) from [<c0287214>] (driver_probe_device+0x48/0x60)
[    0.697692] [<c0287214>] (driver_probe_device+0x48/0x60) from [<c0285af8>] (bus_for_each_drv+0x5c/0x88)
[    0.707458] [<c0285af8>] (bus_for_each_drv+0x5c/0x88) from [<c028718c>] (device_attach+0x98/0xbc)
[    0.716674] [<c028718c>] (device_attach+0x98/0xbc) from [<c0286844>] (bus_probe_device+0x88/0xac)
[    0.725921] [<c0286844>] (bus_probe_device+0x88/0xac) from [<c0285160>] (device_add+0x278/0x358)
[    0.735076] [<c0285160>] (device_add+0x278/0x358) from [<c0327410>] (i2c_new_device+0xec/0x160)
[    0.744354] [<c0327410>] (i2c_new_device+0xec/0x160) from [<c032781c>] (i2c_register_adapter+0x168/0x220)
[    0.754302] [<c032781c>] (i2c_register_adapter+0x168/0x220) from [<c0327a40>] (i2c_add_numbered_adapter+0xd4/0xf0)
[    0.765045] [<c0327a40>] (i2c_add_numbered_adapter+0xd4/0xf0) from [<c041f4c8>] (omap_i2c_probe+0x334/0x424)
[    0.775268] [<c041f4c8>] (omap_i2c_probe+0x334/0x424) from [<c028832c>] (platform_drv_probe+0x18/0x1c)
[    0.784942] [<c028832c>] (platform_drv_probe+0x18/0x1c) from [<c0286fd4>] (really_probe+0x60/0x15c)
[    0.794342] [<c0286fd4>] (really_probe+0x60/0x15c) from [<c0287214>] (driver_probe_device+0x48/0x60)
[    0.803863] [<c0287214>] (driver_probe_device+0x48/0x60) from [<c02872c0>] (__driver_attach+0x94/0x98)
[    0.813537] [<c02872c0>] (__driver_attach+0x94/0x98) from [<c0285b74>] (bus_for_each_dev+0x50/0x7c)
[    0.822937] [<c0285b74>] (bus_for_each_dev+0x50/0x7c) from [<c0286ad4>] (bus_add_driver+0x184/0x248)
[    0.832427] [<c0286ad4>] (bus_add_driver+0x184/0x248) from [<c0287798>] (driver_register+0x78/0x12c)
[    0.841949] [<c0287798>] (driver_register+0x78/0x12c) from [<c0008800>] (do_one_initcall+0x34/0x178)
[    0.851440] [<c0008800>] (do_one_initcall+0x34/0x178) from [<c06068ac>] (kernel_init+0x8c/0x130)
[    0.860565] [<c06068ac>] (kernel_init+0x8c/0x130) from [<c000e850>] (kernel_thread_exit+0x0/0x8)
[    0.870117] ---[ end trace 1b75b31a2719ed1c ]---

Hi Kevin,

Quoted 10 lines · Kevin Hilman

Hi Benoit,

Benoit Cousson<[email protected]> writes:

[...]

Quoted 2 lines

This is not for pull-request, becasue it is based on irqdomain + OMAP IRQ DT series + OMAP twl DT series yet to be pushed.

What's the status of this series?

I sent earlier today the pull request to Samuel.

Here is the branch for information: for_3.4/twl_irq_gpio_mmc_fix

Quoted 3 lines

At least 'gpio/twl: Allocate irq_desc dynamically for SPARSE_IRQ support' is needed in mainline (linux-next) now since without it, non-DT boots will dump a big fat warning trace[2].

Yeah, I know :-(

Quoted 1 line

Also FYI, I needed the patch below[1] in order to build this series.

That's weird, this include was added by the following commit a7cbb9b15d55dff0488b1a6d93929c2386d8632b at 3.2 time frame.

What branch are you using?