Skip to contentSkip to conversation

pinctrl: Introduce PINCTRL_STATE_DEFAULT, define hogs as that state

PATCH In review Stephen Warren 3 participants 11 messages Feb 27, 2012 – Feb 29, 2012

Subsystems (5)
lore
Share & tools
lore
#0 lore raw

PATCH v2 1/2 pinctrl: Introduce PINCTRL_STATE_DEFAULT, define hogs as that state

This provides a single centralized name for the default state.

Update PIN_MAP_* macros to use this state name, instead of requiring the user to pass a state name in.

With this change, hog entries in the mapping table are defined as those with state name PINCTRL_STATE_DEFAULT, i.e. all entries have the same name. This interacts badly with the nested iteration over mapping table entries in pinctrl_hog_maps() and pinctrl_hog_map() which would now attempt to claim each hog mapping table entry multiple times. Replacing the custom hog code with a simple pinctrl_get()/pinctrl_enable().

Update documentation and mapping tables to use this.

---
v2: Added all the hogging rework now described above; hopefully will solve U300 runtime issues.

---
 Documentation/pinctrl.txt       |    8 +-
 arch/arm/mach-u300/core.c       |    6 +-
 drivers/pinctrl/core.c          |  145 ++-------------------------------------
 drivers/pinctrl/core.h          |    4 +-
 include/linux/pinctrl/machine.h |   13 ++--
 include/linux/pinctrl/pinctrl.h |    2 +
 6 files changed, 26 insertions(+), 152 deletions(-)
Documentation/pinctrl.txt +4 -4
index 5e314ce..6fe3232 100644
@@ -814,7 +814,7 @@ it even more compact which assumes you want to use pinctrl-foo and position
 0 for mapping, for example:
 
 static struct pinctrl_map __initdata mapping[] = {
-       PIN_MAP("I2CMAP", "pinctrl-foo", "i2c0", "foo-i2c.0"),
+	PIN_MAP(PINCTRL_STATE_DEFAULT, "pinctrl-foo", "i2c0", "foo-i2c.0"),
 };
 
 
@@ -930,7 +930,7 @@ foo_probe()
 	/* Allocate a state holder named "state" etc */
 	struct pinctrl p;
 
-	p = pinctrl_get(&device, NULL);
+	p = pinctrl_get(&device, PINCTRL_STATE_DEFAULT);
 	if IS_ERR(p)
 		return PTR_ERR(p);
 	pinctrl_enable(p);
@@ -989,7 +989,7 @@ of the pin controller itself, like this:
 
 {
 	.dev_name = "pinctrl-foo",
-	.name = "POWERMAP"
+	.name = PINCTRL_STATE_DEFAULT,
 	.ctrl_dev_name = "pinctrl-foo",
 	.function = "power_func",
 },
@@ -998,7 +998,7 @@ Since it may be common to request the core to hog a few always-applicable
 mux settings on the primary pin controller, there is a convenience macro for
 this:
 
-PIN_MAP_PRIMARY_SYS_HOG("POWERMAP", "pinctrl-foo", "power_func")
+PIN_MAP_SYS_HOG("pinctrl-foo", "power_func")
 
 This gives the exact same result as the above construction.
 
arch/arm/mach-u300/core.c +3 -3
index 336526a..c9050dd 100644
@@ -1569,9 +1569,9 @@ static struct platform_device dma_device = {
 /* Pinmux settings */
 static struct pinctrl_map __initdata u300_pinmux_map[] = {
 	/* anonymous maps for chip power and EMIFs */
-	PIN_MAP_SYS_HOG("POWER", "pinctrl-u300", "power"),
-	PIN_MAP_SYS_HOG("EMIF0", "pinctrl-u300", "emif0"),
-	PIN_MAP_SYS_HOG("EMIF1", "pinctrl-u300", "emif1"),
+	PIN_MAP_SYS_HOG("pinctrl-u300", "power"),
+	PIN_MAP_SYS_HOG("pinctrl-u300", "emif0"),
+	PIN_MAP_SYS_HOG("pinctrl-u300", "emif1"),
 	/* per-device maps for MMC/SD, SPI and UART */
 	PIN_MAP("MMCSD", "pinctrl-u300", "mmc0", "mmci"),
 	PIN_MAP("SPI", "pinctrl-u300", "spi0", "pl022"),
drivers/pinctrl/core.c +7 -138

193 lines not on this page · raw message

drivers/pinctrl/core.h +2 -2
index 7551611..2981b73 100644
@@ -27,7 +27,7 @@ struct pinctrl_gpio_range;
  * @owner: module providing the pin controller, used for refcounting
  * @driver_data: driver data for drivers registering to the pin controller
  *	subsystem
- * @pinctrl_hogs: list of pin control maps hogged by this device
+ * @p: result of pinctrl_get() for this device
  * @device_root: debugfs root for this device
  */
 struct pinctrl_dev {
@@ -39,7 +39,7 @@ struct pinctrl_dev {
 	struct device *dev;
 	struct module *owner;
 	void *driver_data;
-	struct list_head pinctrl_hogs;
+	struct pinctrl *p;
 #ifdef CONFIG_DEBUG_FS
 	struct dentry *device_root;
 #endif
include/linux/pinctrl/machine.h +8 -5
index 73fbb27..20e9735 100644
@@ -12,6 +12,8 @@
 #ifndef __LINUX_PINCTRL_MACHINE_H
 #define __LINUX_PINCTRL_MACHINE_H
 
+#include "pinctrl.h"
+
 /**
  * struct pinctrl_map - boards/machines shall provide this map for devices
  * @dev_name: the name of the device using this specific mapping, the name
@@ -49,17 +51,18 @@ struct pinctrl_map {
  * Convenience macro to map a system function onto a certain pinctrl device,
  * to be hogged by the pin control core until the system shuts down.
  */
-#define PIN_MAP_SYS_HOG(a, b, c) \
-	{ .name = a, .ctrl_dev_name = b, .dev_name = b, .function = c, }
+#define PIN_MAP_SYS_HOG(a, b) \
+	{ .name = PINCTRL_STATE_DEFAULT, .ctrl_dev_name = a, .dev_name = a, \
+	  .function = b, }
 
 /*
  * Convenience macro to map a system function onto a certain pinctrl device
  * using a specified group, to be hogged by the pin control core until the
  * system shuts down.
  */
-#define PIN_MAP_SYS_HOG_GROUP(a, b, c, d)		\
-	{ .name = a, .ctrl_dev_name = b, .dev_name = b, .function = c, \
-	  .group = d, }
+#define PIN_MAP_SYS_HOG_GROUP(a, b, c) \
+	{ .name = PINCTRL_STATE_DEFAULT, .ctrl_dev_name = a, .dev_name = a, \
+	  .function = b, .group = c, }
 
 #ifdef CONFIG_PINMUX
 
include/linux/pinctrl/pinctrl.h +2 -0
index 8bd22ee..411fe23 100644
@@ -19,6 +19,8 @@
 #include <linux/list.h>
 #include <linux/seq_file.h>
 
+#define PINCTRL_STATE_DEFAULT "default"
+
 struct pinctrl_dev;
 struct pinmux_ops;
 struct pinconf_ops;
Signature · 1 line
1.7.0.4
#1 lore raw

PATCH v2 2/2 pinctrl: Assume map table entries can't have a NULL name field

pinctrl_register_mappings() already requires that every mapping table entry have a non-NULL name field.

Logically, this makes sense too; drivers should always request a specific named state so they know what they're getting. Relying on getting the first mentioned state in the mapping table is error-prone, and a nasty special case to implement, given that a given the mapping table may define multiple states for a device.

Remove a small part of the documentation that talked about optionally requesting a specific state; it's mandatory now.

---
v2: Update mach-u300/core.c and sirfsoc_uart.c not to request NULL state names. Use PINCTRL_STATE_DEFAULT from previous patch.

This is now conceptually ack'd by Dong Aisheng, although I didn't actually include the ack above since I've reworked the patch a little since I last posted it (per the v2 changelog above).

---
 Documentation/pinctrl.txt         |    7 +++----
 arch/arm/mach-u300/core.c         |    8 ++++----
 drivers/pinctrl/core.c            |   17 +++++------------
 drivers/tty/serial/sirfsoc_uart.c |    2 +-
 4 files changed, 13 insertions(+), 21 deletions(-)
Documentation/pinctrl.txt +3 -4
index 6fe3232..558aac5 100644
@@ -782,16 +782,19 @@ spi on the second function mapping:
 static const struct pinctrl_map __initdata mapping[] = {
 	{
 		.dev_name = "foo-spi.0",
+		.name = PINCTRL_STATE_DEFAULT,
 		.ctrl_dev_name = "pinctrl-foo",
 		.function = "spi0",
 	},
 	{
 		.dev_name = "foo-i2c.0",
+		.name = PINCTRL_STATE_DEFAULT,
 		.ctrl_dev_name = "pinctrl-foo",
 		.function = "i2c0",
 	},
 	{
 		.dev_name = "foo-mmc.0",
+		.name = PINCTRL_STATE_DEFAULT,
 		.ctrl_dev_name = "pinctrl-foo",
 		.function = "mmc0",
 	},
@@ -944,10 +947,6 @@ foo_remove()
 	pinctrl_put(state->p);
 }
 
-If you want to grab a specific control mapping and not just the first one
-found for this device you can specify a specific mapping name, for example in
-the above example the second i2c0 setting: pinctrl_get(&device, "spi0-pos-B");
-
 This get/enable/disable/put sequence can just as well be handled by bus drivers
 if you don't want each and every driver to handle it and you know the
 arrangement on your bus.
arch/arm/mach-u300/core.c +4 -4
index c9050dd..5b37d84 100644
@@ -1573,9 +1573,9 @@ static struct pinctrl_map __initdata u300_pinmux_map[] = {
 	PIN_MAP_SYS_HOG("pinctrl-u300", "emif0"),
 	PIN_MAP_SYS_HOG("pinctrl-u300", "emif1"),
 	/* per-device maps for MMC/SD, SPI and UART */
-	PIN_MAP("MMCSD", "pinctrl-u300", "mmc0", "mmci"),
-	PIN_MAP("SPI", "pinctrl-u300", "spi0", "pl022"),
-	PIN_MAP("UART0", "pinctrl-u300", "uart0", "uart0"),
+	PIN_MAP(PINCTRL_STATE_DEFAULT, "pinctrl-u300", "mmc0", "mmci"),
+	PIN_MAP(PINCTRL_STATE_DEFAULT, "pinctrl-u300", "spi0", "pl022"),
+	PIN_MAP(PINCTRL_STATE_DEFAULT, "pinctrl-u300", "uart0", "uart0"),
 };
 
 struct u300_mux_hog {
@@ -1607,7 +1607,7 @@ static int __init u300_pinctrl_fetch(void)
 		struct pinctrl *p;
 		int ret;
 
-		p = pinctrl_get(u300_mux_hogs[i].dev, NULL);
+		p = pinctrl_get(u300_mux_hogs[i].dev, PINCTRL_STATE_DEFAULT);
 		if (IS_ERR(p)) {
 			pr_err("u300: could not get pinmux hog %s\n",
 			       u300_mux_hogs[i].name);
drivers/pinctrl/core.c +5 -12
index f25307b..6af6d8d 100644
@@ -461,8 +461,8 @@ static struct pinctrl *pinctrl_get_locked(struct device *dev, const char *name)
 	int i;
 	struct pinctrl_map const *map;
 
-	/* We must have a dev name */
-	if (WARN_ON(!dev))
+	/* We must have both a dev and state name */
+	if (WARN_ON(!dev || !name))
 		return ERR_PTR(-EINVAL);
 
 	devname = dev_name(dev);
@@ -504,16 +504,9 @@ static struct pinctrl *pinctrl_get_locked(struct device *dev, const char *name)
 		if (strcmp(map->dev_name, devname))
 			continue;
 
-		/*
-		 * If we're looking for a specific named map, this must match,
-		 * else we loop and look for the next.
-		 */
-		if (name != NULL) {
-			if (map->name == NULL)
-				continue;
-			if (strcmp(map->name, name))
-				continue;
-		}
+		/* State name must be the one we're looking for */
+		if (strcmp(map->name, name))
+			continue;
 
 		ret = pinmux_apply_muxmap(pctldev, p, dev, devname, map);
 		if (ret) {
drivers/tty/serial/sirfsoc_uart.c +1 -1
index c1a871e..3cabb65 100644
@@ -673,7 +673,7 @@ int sirfsoc_uart_probe(struct platform_device *pdev)
 	port->irq = res->start;
 
 	if (sirfport->hw_flow_ctrl) {
-		sirfport->p = pinctrl_get(&pdev->dev, NULL);
+		sirfport->p = pinctrl_get(&pdev->dev, PINCTRL_STATE_DEFAULT);
 		ret = IS_ERR(sirfport->p);
 		if (ret)
 			goto pin_err;
Signature · 1 line
1.7.0.4
Acked-by on v2 2/2
#2 lore raw

Re: [PATCH V2 2/2] pinctrl: Assume map table entries can't have a NULL name field

Quoted 13 lines · Stephen Warren

pinctrl_register_mappings() already requires that every mapping table entry have a non-NULL name field.

Logically, this makes sense too; drivers should always request a specific named state so they know what they're getting. Relying on getting the first mentioned state in the mapping table is error-prone, and a nasty special case to implement, given that a given the mapping table may define multiple states for a device.

Remove a small part of the documentation that talked about optionally requesting a specific state; it's mandatory now.

Quoted 8 lines

---
v2: Update mach-u300/core.c and sirfsoc_uart.c not to request NULL state names. Use PINCTRL_STATE_DEFAULT from previous patch.

This is now conceptually ack'd by Dong Aisheng, although I didn't actually include the ack above since I've reworked the patch a little since I last posted it (per the v2 changelog above).
---

It looks very good to me.

Regards
Dong Aisheng

Quoted 119 lines
 Documentation/pinctrl.txt         |    7 +++----
 arch/arm/mach-u300/core.c         |    8 ++++----
 drivers/pinctrl/core.c            |   17 +++++------------
 drivers/tty/serial/sirfsoc_uart.c |    2 +-
 4 files changed, 13 insertions(+), 21 deletions(-)
Documentation/pinctrl.txt +3 -4
index 6fe3232..558aac5 100644
@@ -782,16 +782,19 @@ spi on the second function mapping:
 static const struct pinctrl_map __initdata mapping[] = {
 	{
 		.dev_name = "foo-spi.0",
+		.name = PINCTRL_STATE_DEFAULT,
 		.ctrl_dev_name = "pinctrl-foo",
 		.function = "spi0",
 	},
 	{
 		.dev_name = "foo-i2c.0",
+		.name = PINCTRL_STATE_DEFAULT,
 		.ctrl_dev_name = "pinctrl-foo",
 		.function = "i2c0",
 	},
 	{
 		.dev_name = "foo-mmc.0",
+		.name = PINCTRL_STATE_DEFAULT,
 		.ctrl_dev_name = "pinctrl-foo",
 		.function = "mmc0",
 	},
@@ -944,10 +947,6 @@ foo_remove()
 	pinctrl_put(state->p);
 }
 
-If you want to grab a specific control mapping and not just the first one
-found for this device you can specify a specific mapping name, for example in
-the above example the second i2c0 setting: pinctrl_get(&device, "spi0-pos-B");
-
 This get/enable/disable/put sequence can just as well be handled by bus drivers
 if you don't want each and every driver to handle it and you know the
 arrangement on your bus.
arch/arm/mach-u300/core.c +4 -4
index c9050dd..5b37d84 100644
@@ -1573,9 +1573,9 @@ static struct pinctrl_map __initdata u300_pinmux_map[] = {
 	PIN_MAP_SYS_HOG("pinctrl-u300", "emif0"),
 	PIN_MAP_SYS_HOG("pinctrl-u300", "emif1"),
 	/* per-device maps for MMC/SD, SPI and UART */
-	PIN_MAP("MMCSD", "pinctrl-u300", "mmc0", "mmci"),
-	PIN_MAP("SPI", "pinctrl-u300", "spi0", "pl022"),
-	PIN_MAP("UART0", "pinctrl-u300", "uart0", "uart0"),
+	PIN_MAP(PINCTRL_STATE_DEFAULT, "pinctrl-u300", "mmc0", "mmci"),
+	PIN_MAP(PINCTRL_STATE_DEFAULT, "pinctrl-u300", "spi0", "pl022"),
+	PIN_MAP(PINCTRL_STATE_DEFAULT, "pinctrl-u300", "uart0", "uart0"),
 };
 
 struct u300_mux_hog {
@@ -1607,7 +1607,7 @@ static int __init u300_pinctrl_fetch(void)
 		struct pinctrl *p;
 		int ret;
 
-		p = pinctrl_get(u300_mux_hogs[i].dev, NULL);
+		p = pinctrl_get(u300_mux_hogs[i].dev, PINCTRL_STATE_DEFAULT);
 		if (IS_ERR(p)) {
 			pr_err("u300: could not get pinmux hog %s\n",
 			       u300_mux_hogs[i].name);
drivers/pinctrl/core.c +5 -12
index f25307b..6af6d8d 100644
@@ -461,8 +461,8 @@ static struct pinctrl *pinctrl_get_locked(struct device *dev, const char *name)
 	int i;
 	struct pinctrl_map const *map;
 
-	/* We must have a dev name */
-	if (WARN_ON(!dev))
+	/* We must have both a dev and state name */
+	if (WARN_ON(!dev || !name))
 		return ERR_PTR(-EINVAL);
 
 	devname = dev_name(dev);
@@ -504,16 +504,9 @@ static struct pinctrl *pinctrl_get_locked(struct device *dev, const char *name)
 		if (strcmp(map->dev_name, devname))
 			continue;
 
-		/*
-		 * If we're looking for a specific named map, this must match,
-		 * else we loop and look for the next.
-		 */
-		if (name != NULL) {
-			if (map->name == NULL)
-				continue;
-			if (strcmp(map->name, name))
-				continue;
-		}
+		/* State name must be the one we're looking for */
+		if (strcmp(map->name, name))
+			continue;
 
 		ret = pinmux_apply_muxmap(pctldev, p, dev, devname, map);
 		if (ret) {
drivers/tty/serial/sirfsoc_uart.c +1 -1
index c1a871e..3cabb65 100644
@@ -673,7 +673,7 @@ int sirfsoc_uart_probe(struct platform_device *pdev)
 	port->irq = res->start;
 
 	if (sirfport->hw_flow_ctrl) {
-		sirfport->p = pinctrl_get(&pdev->dev, NULL);
+		sirfport->p = pinctrl_get(&pdev->dev, PINCTRL_STATE_DEFAULT);
 		ret = IS_ERR(sirfport->p);
 		if (ret)
 			goto pin_err;
Signature · 1 line
1.7.0.4
Acked-by on v2 1/2
#3 lore raw
In reply to Stephen Warren
Quoted 12 lines · Stephen Warren

This provides a single centralized name for the default state.

Update PIN_MAP_* macros to use this state name, instead of requiring the user to pass a state name in.

With this change, hog entries in the mapping table are defined as those with state name PINCTRL_STATE_DEFAULT, i.e. all entries have the same name. This interacts badly with the nested iteration over mapping table entries in pinctrl_hog_maps() and pinctrl_hog_map() which would now attempt to claim each hog mapping table entry multiple times. Replacing the custom hog code with a simple pinctrl_get()/pinctrl_enable().

Quoted 4 lines

Update documentation and mapping tables to use this.

---

The whole patch looks good to me.
A few trivial comments:

Quoted 16 lines

v2: Added all the hogging rework now described above; hopefully will solve U300 runtime issues.

---
 Documentation/pinctrl.txt       |    8 +-
 arch/arm/mach-u300/core.c       |    6 +-
 drivers/pinctrl/core.c          |  145 ++-------------------------------------
 drivers/pinctrl/core.h          |    4 +-
 include/linux/pinctrl/machine.h |   13 ++--
 include/linux/pinctrl/pinctrl.h |    2 +
 6 files changed, 26 insertions(+), 152 deletions(-)
Documentation/pinctrl.txt +0 -0
index 5e314ce..6fe3232 100644

Documentation/pinctrl.txt it even more compact which assumes you want to use pinctrl-foo and position · line 817

Quoted 5 lines
Documentation/pinctrl.txt +1 -1
@@ -814,7 +814,7 @@ it even more compact which assumes you want to use pinctrl-foo and position
 0 for mapping, for example:
 
 static struct pinctrl_map __initdata mapping[] = {
-       PIN_MAP("I2CMAP", "pinctrl-foo", "i2c0", "foo-i2c.0"),
+	PIN_MAP(PINCTRL_STATE_DEFAULT, "pinctrl-foo", "i2c0", "foo-i2c.0"),

It seemed Linus already applied your "re-order struct pinctrl_map". This may need change a bit according to that patch.

Quoted 12 lines
Documentation/pinctrl.txt +1 -1
@@ -930,7 +930,7 @@ foo_probe()
 	/* Allocate a state holder named "state" etc */
 	struct pinctrl p;
 
-	p = pinctrl_get(&device, NULL);
+	p = pinctrl_get(&device, PINCTRL_STATE_DEFAULT);
 	if IS_ERR(p)
 		return PTR_ERR(p);
 	pinctrl_enable(p);

Documentation/pinctrl.txt of the pin controller itself, like this: · line 995

Quoted 8 lines
Documentation/pinctrl.txt +1 -1
@@ -989,7 +989,7 @@ of the pin controller itself, like this:
 
 {
 	.dev_name = "pinctrl-foo",
-	.name = "POWERMAP"
+	.name = PINCTRL_STATE_DEFAULT,
 	.ctrl_dev_name = "pinctrl-foo",
 	.function = "power_func",
 },

ditto

Quoted 200 lines
Documentation/pinctrl.txt +1 -1
@@ -998,7 +998,7 @@ Since it may be common to request the core to hog a few always-applicable
 mux settings on the primary pin controller, there is a convenience macro for
 this:
 
-PIN_MAP_PRIMARY_SYS_HOG("POWERMAP", "pinctrl-foo", "power_func")
+PIN_MAP_SYS_HOG("pinctrl-foo", "power_func")
 
 This gives the exact same result as the above construction.
 
arch/arm/mach-u300/core.c +3 -3
index 336526a..c9050dd 100644
@@ -1569,9 +1569,9 @@ static struct platform_device dma_device = {
 /* Pinmux settings */
 static struct pinctrl_map __initdata u300_pinmux_map[] = {
 	/* anonymous maps for chip power and EMIFs */
-	PIN_MAP_SYS_HOG("POWER", "pinctrl-u300", "power"),
-	PIN_MAP_SYS_HOG("EMIF0", "pinctrl-u300", "emif0"),
-	PIN_MAP_SYS_HOG("EMIF1", "pinctrl-u300", "emif1"),
+	PIN_MAP_SYS_HOG("pinctrl-u300", "power"),
+	PIN_MAP_SYS_HOG("pinctrl-u300", "emif0"),
+	PIN_MAP_SYS_HOG("pinctrl-u300", "emif1"),
 	/* per-device maps for MMC/SD, SPI and UART */
 	PIN_MAP("MMCSD", "pinctrl-u300", "mmc0", "mmci"),
 	PIN_MAP("SPI", "pinctrl-u300", "spi0", "pl022"),
drivers/pinctrl/core.c +0 -134
index 376cede..f25307b 100644
@@ -44,18 +44,6 @@ struct pinctrl_maps {
 	unsigned num_maps;
 };
 
-/**
- * struct pinctrl_hog - a list item to stash control hogs
- * @node: pin control hog list node
- * @map: map entry responsible for this hogging
- * @pmx: the pin control hogged by this item
- */
-struct pinctrl_hog {
-	struct list_head node;
-	struct pinctrl_map const *map;
-	struct pinctrl *p;
-};
-
 /* Global list of pin control devices */
 static DEFINE_MUTEX(pinctrldev_list_mutex);
 static LIST_HEAD(pinctrldev_list);
@@ -702,103 +690,6 @@ int pinctrl_register_mappings(struct pinctrl_map const *maps,
 	return 0;
 }
 
-/* Hog a single map entry and add to the hoglist */
-static int pinctrl_hog_map(struct pinctrl_dev *pctldev,
-			   struct pinctrl_map const *map)
-{
-	struct pinctrl_hog *hog;
-	struct pinctrl *p;
-	int ret;
-
-	hog = kzalloc(sizeof(*hog), GFP_KERNEL);
-	if (!hog) {
-		dev_err(pctldev->dev, "failed to alloc struct pinctrl_hog\n");
-		return -ENOMEM;
-	}
-
-	p = pinctrl_get_locked(pctldev->dev, map->name);
-	if (IS_ERR(p)) {
-		kfree(hog);
-		dev_err(pctldev->dev,
-			"could not get the %s pin control mapping for hogging\n",
-			map->name);
-		return PTR_ERR(p);
-	}
-
-	ret = pinctrl_enable(p);
-	if (ret) {
-		pinctrl_put(p);
-		kfree(hog);
-		dev_err(pctldev->dev,
-			"could not enable the %s pin control mapping for hogging\n",
-			map->name);
-		return ret;
-	}
-
-	hog->map = map;
-	hog->p = p;
-
-	dev_info(pctldev->dev, "hogged map %s, function %s\n", map->name,
-		 map->function);
-	list_add_tail(&hog->node, &pctldev->pinctrl_hogs);
-
-	return 0;
-}
-
-/**
- * pinctrl_hog_maps() - hog specific map entries on controller device
- * @pctldev: the pin control device to hog entries on
- *
- * When the pin controllers are registered, there may be some specific pinmux
- * map entries that need to be hogged, i.e. get+enabled until the system shuts
- * down.
- */
-static int pinctrl_hog_maps(struct pinctrl_dev *pctldev)
-{
-	struct device *dev = pctldev->dev;
-	const char *devname = dev_name(dev);
-	int ret;
-	struct pinctrl_maps *maps_node;
-	int i;
-	struct pinctrl_map const *map;
-
-	INIT_LIST_HEAD(&pctldev->pinctrl_hogs);
-
-	mutex_lock(&pinctrl_maps_mutex);
-	for_each_maps(maps_node, i, map) {
-		if (!strcmp(map->ctrl_dev_name, devname) &&
-		    !strcmp(map->dev_name, devname)) {
-			/* OK time to hog! */
-			ret = pinctrl_hog_map(pctldev, map);
-			if (ret) {
-				mutex_unlock(&pinctrl_maps_mutex);
-				return ret;
-			}
-		}
-	}
-	mutex_unlock(&pinctrl_maps_mutex);
-
-	return 0;
-}
-
-/**
- * pinctrl_unhog_maps() - unhog specific map entries on controller device
- * @pctldev: the pin control device to unhog entries on
- */
-static void pinctrl_unhog_maps(struct pinctrl_dev *pctldev)
-{
-	struct list_head *node, *tmp;
-
-	list_for_each_safe(node, tmp, &pctldev->pinctrl_hogs) {
-		struct pinctrl_hog *hog =
-			list_entry(node, struct pinctrl_hog, node);
-		pinctrl_disable(hog->p);
-		pinctrl_put(hog->p);
-		list_del(node);
-		kfree(hog);
-	}
-}
-
 #ifdef CONFIG_DEBUG_FS
 
 static int pinctrl_pins_show(struct seq_file *s, void *what)
@@ -889,19 +780,6 @@ static int pinctrl_gpioranges_show(struct seq_file *s, void *what)
 	return 0;
 }
 
-static int pinmux_hogs_show(struct seq_file *s, void *what)
-{
-	struct pinctrl_dev *pctldev = s->private;
-	struct pinctrl_hog *hog;
-
-	seq_puts(s, "Pin control map hogs held by device\n");
-
-	list_for_each_entry(hog, &pctldev->pinctrl_hogs, node)
-		seq_printf(s, "%s\n", hog->map->name);
-
-	return 0;
-}
-
 static int pinctrl_devices_show(struct seq_file *s, void *what)
 {
 	struct pinctrl_dev *pctldev;
@@ -988,11 +866,6 @@ static int pinctrl_gpioranges_open(struct inode *inode, struct file *file)
 	return single_open(file, pinctrl_gpioranges_show, inode->i_private);
 }
 
-static int pinmux_hogs_open(struct inode *inode, struct file *file)
-{
-	return single_open(file, pinmux_hogs_show, inode->i_private);
-}
-
 static int pinctrl_devices_open(struct inode *inode, struct file *file)
 {
 	return single_open(file, pinctrl_devices_show, NULL);
@@ -1029,13 +902,6 @@ static const struct file_operations pinctrl_gpioranges_ops = {
 	.release	= single_release,
 };
 
-static const struct file_operations pinmux_hogs_ops = {
-	.open		= pinmux_hogs_open,
-	.read		= seq_read,
-	.llseek		= seq_lseek,
-	.release	= single_release,
-};
-
 static const struct file_operations pinctrl_devices_ops = {
 	.open		= pinctrl_devices_open,
 	.read		= seq_read,

drivers/pinctrl/core.c static void pinctrl_init_device_debugfs(struct pinctrl_dev *pctldev) · line 1082

Quoted 5 lines
drivers/pinctrl/core.c +0 -2
@@ -1078,8 +944,6 @@ static void pinctrl_init_device_debugfs(struct pinctrl_dev *pctldev)
 			    device_root, pctldev, &pinctrl_groups_ops);
 	debugfs_create_file("gpio-ranges", S_IFREG | S_IRUGO,
 			    device_root, pctldev, &pinctrl_gpioranges_ops);
-	debugfs_create_file("pinmux-hogs", S_IFREG | S_IRUGO,
-			    device_root, pctldev, &pinmux_hogs_ops);

I see you remove the pinnmux-hogs sysfs file here.
I guess you may want to merge it into pinctrl-maps sysfs file since they're almost same(only difference is device name).

Do you think if we can add a special flag for that type of map in sysfs (e.g. a "Hog" flag behind the regular map debug info)?
Then users do not need to check device's name to see if it's a hogged function.

(Anyway, it should be in another patch)

drivers/pinctrl/core.c struct pinctrl_dev *pinctrl_register(struct pinctrl_desc *pctldesc, · line 1057

Quoted 11 lines
drivers/pinctrl/core.c +3 -1
 	pinmux_init_device_debugfs(device_root, pctldev);
 	pinconf_init_device_debugfs(device_root, pctldev);
 }
@@ -1188,7 +1052,9 @@ struct pinctrl_dev *pinctrl_register(struct pinctrl_desc *pctldesc,
 	mutex_lock(&pinctrldev_list_mutex);
 	list_add_tail(&pctldev->node, &pinctrldev_list);
 	mutex_unlock(&pinctrldev_list_mutex);
-	pinctrl_hog_maps(pctldev);
+	pctldev->p = pinctrl_get(pctldev->dev, PINCTRL_STATE_DEFAULT);
+	if (!IS_ERR(pctldev->p))
+		pinctrl_enable(pctldev->p);

This change looks good to me.

Quoted 62 lines
drivers/pinctrl/core.c +4 -1
 	pinctrl_init_device_debugfs(pctldev);
 
 	return pctldev;
@@ -1211,7 +1077,10 @@ void pinctrl_unregister(struct pinctrl_dev *pctldev)
 		return;
 
 	pinctrl_remove_device_debugfs(pctldev);
-	pinctrl_unhog_maps(pctldev);
+	if (!IS_ERR(pctldev->p)) {
+		pinctrl_disable(pctldev->p);
+		pinctrl_put(pctldev->p);
+	}
 	/* TODO: check that no pinmuxes are still active? */
 	mutex_lock(&pinctrldev_list_mutex);
 	list_del(&pctldev->node);
drivers/pinctrl/core.h +2 -2
index 7551611..2981b73 100644
@@ -27,7 +27,7 @@ struct pinctrl_gpio_range;
  * @owner: module providing the pin controller, used for refcounting
  * @driver_data: driver data for drivers registering to the pin controller
  *	subsystem
- * @pinctrl_hogs: list of pin control maps hogged by this device
+ * @p: result of pinctrl_get() for this device
  * @device_root: debugfs root for this device
  */
 struct pinctrl_dev {
@@ -39,7 +39,7 @@ struct pinctrl_dev {
 	struct device *dev;
 	struct module *owner;
 	void *driver_data;
-	struct list_head pinctrl_hogs;
+	struct pinctrl *p;
 #ifdef CONFIG_DEBUG_FS
 	struct dentry *device_root;
 #endif
include/linux/pinctrl/machine.h +5 -2
index 73fbb27..20e9735 100644
@@ -12,6 +12,8 @@
 #ifndef __LINUX_PINCTRL_MACHINE_H
 #define __LINUX_PINCTRL_MACHINE_H
 
+#include "pinctrl.h"
+
 /**
  * struct pinctrl_map - boards/machines shall provide this map for devices
  * @dev_name: the name of the device using this specific mapping, the name
@@ -49,17 +51,18 @@ struct pinctrl_map {
  * Convenience macro to map a system function onto a certain pinctrl device,
  * to be hogged by the pin control core until the system shuts down.
  */
-#define PIN_MAP_SYS_HOG(a, b, c) \
-	{ .name = a, .ctrl_dev_name = b, .dev_name = b, .function = c, }
+#define PIN_MAP_SYS_HOG(a, b) \
+	{ .name = PINCTRL_STATE_DEFAULT, .ctrl_dev_name = a, .dev_name = a, \
+	  .function = b, }
 
 /*
  * Convenience macro to map a system function onto a certain pinctrl device

include/linux/pinctrl/machine.h struct pinctrl_map { · line 66

Quoted 10 lines
include/linux/pinctrl/machine.h +3 -3
  * using a specified group, to be hogged by the pin control core until the
  * system shuts down.
  */
-#define PIN_MAP_SYS_HOG_GROUP(a, b, c, d)		\
-	{ .name = a, .ctrl_dev_name = b, .dev_name = b, .function = c, \
-	  .group = d, }
+#define PIN_MAP_SYS_HOG_GROUP(a, b, c) \
+	{ .name = PINCTRL_STATE_DEFAULT, .ctrl_dev_name = a, .dev_name = a, \
+	  .function = b, .group = c, }
 

In your v1 patch, we discussed about the possible requirement of different state support for hog functions(it seemed still no conclusion, right?). If we decide to support it, i'd suggest change here macro a bit to align with the exist PIN_MAP macro like:

#define PIN_MAP_SYS_HOG_DEFAULT(a, b) \
	{ .name = PINCTRL_STATE_DEFAULT, .ctrl_dev_name = a, .dev_name = a, \
	.function = b, }
#define PIN_MAP_SYS_HOG_GROUP_DEFAULT(a, b, c) \
	{ .name = PINCTRL_STATE_DEFAULT, .ctrl_dev_name = a, .dev_name = a, \
	.function = b, .group = c, }

And reserve the original PIN_MAP_SYS_HOG for users to add the state name as they want.

However, since this is still undetermined, i'm also ok if you want to use this macro first.

Regards
Dong Aisheng

#4 lore raw

Dong Aisheng wrote at Tuesday, February 28, 2012 2:30 AM:

Quoted 5 lines · Stephen Warren
Quoted 4 lines

This provides a single centralized name for the default state.

Update PIN_MAP_* macros to use this state name, instead of requiring the user to pass a state name in.

...

Quoted 8 lines

Documentation/pinctrl.txt it even more compact which assumes you want to use pinctrl-foo and position · line 817

Quoted 5 lines
Documentation/pinctrl.txt +1 -1
 
 static struct pinctrl_map __initdata mapping[] = {
-       PIN_MAP("I2CMAP", "pinctrl-foo", "i2c0", "foo-i2c.0"),
+	PIN_MAP(PINCTRL_STATE_DEFAULT, "pinctrl-foo", "i2c0", "foo-i2c.0"),

It seemed Linus already applied your "re-order struct pinctrl_map". This may need change a bit according to that patch.

Yes, that patch is applied.

However, note that I deliberately did not change PIN_MAP()'s parameter order in that patch. I deferred as much rework of those macros to the later patch "pinctrl: Enhance mapping table to support pin config operations" to avoid repeatedly changing those macros where possible.

For reference, the definition of PIN_MAP() at this point in the series is:

#define PIN_MAP(a, b, c, d) \
        { .name = a, .ctrl_dev_name = b, .function = c, .dev_name = d }
Quoted 11 lines

Documentation/pinctrl.txt of the pin controller itself, like this: · line 995

Quoted 9 lines
Documentation/pinctrl.txt +1 -1
@@ -989,7 +989,7 @@ of the pin controller itself, like this:
 
 {
 	.dev_name = "pinctrl-foo",
-	.name = "POWERMAP"
+	.name = PINCTRL_STATE_DEFAULT,
 	.ctrl_dev_name = "pinctrl-foo",
 	.function = "power_func",
 },

ditto

That already matches the order of fields in the struct definition, which for reference is as follows at this point in the series:

struct pinctrl_map {
        const char *dev_name;
        const char *name;
        const char *ctrl_dev_name;
        const char *function;
        const char *group;
};

(.group is still allowed to be NULL)

Quoted 11 lines

Documentation/pinctrl.txt static void pinctrl_init_device_debugfs(struct pinctrl_dev *pctldev) · line 1082

Quoted 6 lines
Documentation/pinctrl.txt +0 -2
@@ -1078,8 +944,6 @@ static void pinctrl_init_device_debugfs(struct pinctrl_dev *pctldev)
 			    device_root, pctldev, &pinctrl_groups_ops);
 	debugfs_create_file("gpio-ranges", S_IFREG | S_IRUGO,
 			    device_root, pctldev, &pinctrl_gpioranges_ops);
-	debugfs_create_file("pinmux-hogs", S_IFREG | S_IRUGO,
-			    device_root, pctldev, &pinmux_hogs_ops);

I see you remove the pinnmux-hogs sysfs file here.
I guess you may want to merge it into pinctrl-maps sysfs file since they're almost same(only difference is device name).

Quoted 5 lines

Do you think if we can add a special flag for that type of map in sysfs (e.g. a "Hog" flag behind the regular map debug info)?
Then users do not need to check device's name to see if it's a hogged function.

(Anyway, it should be in another patch)

I'd rather not myself; I don't see why "hog" should be a special-case; it's just that the pin controller driver's mapping table entries have it set up certain pin mux/config settings.

That said, feel free to submit a patch for this. I'd prefer that any text you add to the debugfs files to indicate this "hogging" be in addition to anything that's already present rather than replacing it (as has unfortunately happened already in "pinmux-pins"), otherwise it obscures information.

Quoted 23 lines
Quoted 23 lines
include/linux/pinctrl/machine.h +5 -2
index 73fbb27..20e9735 100644
@@ -12,6 +12,8 @@
 #ifndef __LINUX_PINCTRL_MACHINE_H
 #define __LINUX_PINCTRL_MACHINE_H
 
+#include "pinctrl.h"
+
 /**
  * struct pinctrl_map - boards/machines shall provide this map for devices
  * @dev_name: the name of the device using this specific mapping, the name
@@ -49,17 +51,18 @@ struct pinctrl_map {
  * Convenience macro to map a system function onto a certain pinctrl device,
  * to be hogged by the pin control core until the system shuts down.
  */
-#define PIN_MAP_SYS_HOG(a, b, c) \
-	{ .name = a, .ctrl_dev_name = b, .dev_name = b, .function = c, }
+#define PIN_MAP_SYS_HOG(a, b) \
+	{ .name = PINCTRL_STATE_DEFAULT, .ctrl_dev_name = a, .dev_name = a, \
+	  .function = b, }
 
Quoted 15 lines

include/linux/pinctrl/machine.h struct pinctrl_map { · line 66

Quoted 12 lines
include/linux/pinctrl/machine.h +3 -3
 /*
  * Convenience macro to map a system function onto a certain pinctrl device
  * using a specified group, to be hogged by the pin control core until the
  * system shuts down.
  */
-#define PIN_MAP_SYS_HOG_GROUP(a, b, c, d)		\
-	{ .name = a, .ctrl_dev_name = b, .dev_name = b, .function = c, \
-	  .group = d, }
+#define PIN_MAP_SYS_HOG_GROUP(a, b, c) \
+	{ .name = PINCTRL_STATE_DEFAULT, .ctrl_dev_name = a, .dev_name = a, \
+	  .function = b, .group = c, }
 

In your v1 patch, we discussed about the possible requirement of different state support for hog functions(it seemed still no conclusion, right?).

I agreed to fully support that. However as I mentioned, the support will be actually implemented in patch "pinctrl: Enhance mapping table to support pin config operations" since I was already reworking all the mapping table macros there, it felt better to do the rework once in that patch.

I'm waiting on resolution of the locking patch issue before I repost an updated version of the mapping table rework patch that includes this support.

Quoted 1 line

Thanks.

Signature · 1 line
nvpublic
#5 lore raw
Quoted 24 lines · Stephen Warren

Dong Aisheng wrote at Tuesday, February 28, 2012 2:30 AM:

Quoted 5 lines · Stephen Warren
Quoted 4 lines

This provides a single centralized name for the default state.

Update PIN_MAP_* macros to use this state name, instead of requiring the user to pass a state name in.

...

Quoted 8 lines

Documentation/pinctrl.txt it even more compact which assumes you want to use pinctrl-foo and position · line 817

Quoted 5 lines
Documentation/pinctrl.txt +1 -1
 
 static struct pinctrl_map __initdata mapping[] = {
-       PIN_MAP("I2CMAP", "pinctrl-foo", "i2c0", "foo-i2c.0"),
+	PIN_MAP(PINCTRL_STATE_DEFAULT, "pinctrl-foo", "i2c0", "foo-i2c.0"),

It seemed Linus already applied your "re-order struct pinctrl_map". This may need change a bit according to that patch.

Yes, that patch is applied.

However, note that I deliberately did not change PIN_MAP()'s parameter order in that patch. I deferred as much rework of those macros to the later patch "pinctrl: Enhance mapping table to support pin config operations" to avoid repeatedly changing those macros where possible.

Quoted 6 lines

For reference, the definition of PIN_MAP() at this point in the series is:

#define PIN_MAP(a, b, c, d) \
        { .name = a, .ctrl_dev_name = b, .function = c, .dev_name = d }

Ok, i see.

Quoted 23 lines

(.group is still allowed to be NULL)

Quoted 11 lines

Documentation/pinctrl.txt static void pinctrl_init_device_debugfs(struct pinctrl_dev *pctldev) · line 1082

Quoted 6 lines
Documentation/pinctrl.txt +0 -2
@@ -1078,8 +944,6 @@ static void pinctrl_init_device_debugfs(struct pinctrl_dev *pctldev)
 			    device_root, pctldev, &pinctrl_groups_ops);
 	debugfs_create_file("gpio-ranges", S_IFREG | S_IRUGO,
 			    device_root, pctldev, &pinctrl_gpioranges_ops);
-	debugfs_create_file("pinmux-hogs", S_IFREG | S_IRUGO,
-			    device_root, pctldev, &pinmux_hogs_ops);

I see you remove the pinnmux-hogs sysfs file here.
I guess you may want to merge it into pinctrl-maps sysfs file since they're almost same(only difference is device name).

Quoted 5 lines

Do you think if we can add a special flag for that type of map in sysfs (e.g. a "Hog" flag behind the regular map debug info)?
Then users do not need to check device's name to see if it's a hogged function.

(Anyway, it should be in another patch)

I'd rather not myself; I don't see why "hog" should be a special-case; it's just that the pin controller driver's mapping table entries have it set up certain pin mux/config settings.

Quoted 6 lines

That said, feel free to submit a patch for this. I'd prefer that any text you add to the debugfs files to indicate this "hogging" be in addition to anything that's already present rather than replacing it (as has unfortunately happened already in "pinmux-pins"), otherwise it obscures information.

Yes, correct.

Quoted 39 lines
Quoted 23 lines
Quoted 23 lines
include/linux/pinctrl/machine.h +5 -2
index 73fbb27..20e9735 100644
@@ -12,6 +12,8 @@
 #ifndef __LINUX_PINCTRL_MACHINE_H
 #define __LINUX_PINCTRL_MACHINE_H
 
+#include "pinctrl.h"
+
 /**
  * struct pinctrl_map - boards/machines shall provide this map for devices
  * @dev_name: the name of the device using this specific mapping, the name
@@ -49,17 +51,18 @@ struct pinctrl_map {
  * Convenience macro to map a system function onto a certain pinctrl device,
  * to be hogged by the pin control core until the system shuts down.
  */
-#define PIN_MAP_SYS_HOG(a, b, c) \
-	{ .name = a, .ctrl_dev_name = b, .dev_name = b, .function = c, }
+#define PIN_MAP_SYS_HOG(a, b) \
+	{ .name = PINCTRL_STATE_DEFAULT, .ctrl_dev_name = a, .dev_name = a, \
+	  .function = b, }
 
Quoted 15 lines

include/linux/pinctrl/machine.h struct pinctrl_map { · line 66

Quoted 12 lines
include/linux/pinctrl/machine.h +3 -3
 /*
  * Convenience macro to map a system function onto a certain pinctrl device
  * using a specified group, to be hogged by the pin control core until the
  * system shuts down.
  */
-#define PIN_MAP_SYS_HOG_GROUP(a, b, c, d)		\
-	{ .name = a, .ctrl_dev_name = b, .dev_name = b, .function = c, \
-	  .group = d, }
+#define PIN_MAP_SYS_HOG_GROUP(a, b, c) \
+	{ .name = PINCTRL_STATE_DEFAULT, .ctrl_dev_name = a, .dev_name = a, \
+	  .function = b, .group = c, }
 

In your v1 patch, we discussed about the possible requirement of different state support for hog functions(it seemed still no conclusion, right?).

Quoted 1 line

I agreed to fully support that. However as I mentioned, the support will be

Great.

Quoted 4 lines

actually implemented in patch "pinctrl: Enhance mapping table to support pin config operations" since I was already reworking all the mapping table macros there, it felt better to do the rework once in that patch.

Reasonable to me.

Quoted 4 lines

I'm waiting on resolution of the locking patch issue before I repost an updated version of the mapping table rework patch that includes this support.

Sounds good.

Quoted 2 lines
Quoted 1 line

Regards
Dong Aisheng

#6 lore raw
In reply to Stephen Warren
Quoted 12 lines · Stephen Warren

This provides a single centralized name for the default state.

Update PIN_MAP_* macros to use this state name, instead of requiring the user to pass a state name in.

With this change, hog entries in the mapping table are defined as those with state name PINCTRL_STATE_DEFAULT, i.e. all entries have the same name. This interacts badly with the nested iteration over mapping table entries in pinctrl_hog_maps() and pinctrl_hog_map() which would now attempt to claim each hog mapping table entry multiple times. Replacing the custom hog code with a simple pinctrl_get()/pinctrl_enable().

Quoted 6 lines

Update documentation and mapping tables to use this.

---
v2: Added all the hogging rework now described above; hopefully will solve U300 runtime issues.

This is a bit better, I now have my devices but I lost my hogs :-(

Trying to see if I can fix it...

Yours,
Linus Walleij

#7 lore raw
In reply to Stephen Warren

Aha now I see the problem here...

Quoted 1 line · Stephen Warren

This provides a single centralized name for the default state.

(...)

Quoted 2 lines
drivers/pinctrl/core.c +0 -2
-/* Hog a single map entry and add to the hoglist */
-static int pinctrl_hog_map(struct pinctrl_dev *pctldev,

(...)

Quoted 1 line

-static int pinctrl_hog_maps(struct pinctrl_dev *pctldev)

(...)

Quoted 4 lines

-       pinctrl_hog_maps(pctldev);
+       pctldev->p = pinctrl_get(pctldev->dev, PINCTRL_STATE_DEFAULT);
+       if (!IS_ERR(pctldev->p))
+               pinctrl_enable(pctldev->p);

So what happens here is that my hogs will try to activate three different functions in the same struct pinctrl *p.

This fails the sanity check in pinmux.c:

	/*
	 * If the function selector is already set, it needs to be identical,
	 * we support several groups with one function but not several
	 * functions with one or several groups in the same pinmux.
	 */
	if (p->func_selector != UINT_MAX &&
	    p->func_selector != func_selector) {
		dev_err(pctldev->dev,
			"dual function defines in the map for device %s\n",
		       devname);
		return -EINVAL;
	}
	p->func_selector = func_selector;

Since it is assumed that we can only activate one function at a time for a given device handle (struct pinmux).

Which was sound, since we said that one single function can activate several groups.

So you still need one struct pinmux *p for each hog or it won't work.

Allowing the struct pinmux to activate several functions at once is another solution, but would defy the idea of having several groups mapped to one function, then you could just as well use that instead, and it would be a tad bit too many "many to many" relations in my opinion.

So I would revert back to using one handle per hog.

(Sorry if I misunderstood the problem though we're getting pretty complex here...)

Yours,
Linus Walleij

#8 lore raw

Linus Walleij wrote at Wednesday, February 29, 2012 9:41 AM:

Quoted 21 lines

Aha now I see the problem here...

Quoted 1 line · Stephen Warren

This provides a single centralized name for the default state.

(...)

Quoted 2 lines
drivers/pinctrl/core.c +0 -2
-/* Hog a single map entry and add to the hoglist */
-static int pinctrl_hog_map(struct pinctrl_dev *pctldev,

(...)

Quoted 1 line

-static int pinctrl_hog_maps(struct pinctrl_dev *pctldev)

(...)

Quoted 4 lines

-       pinctrl_hog_maps(pctldev);
+       pctldev->p = pinctrl_get(pctldev->dev, PINCTRL_STATE_DEFAULT);
+       if (!IS_ERR(pctldev->p))
+               pinctrl_enable(pctldev->p);

So what happens here is that my hogs will try to activate three different functions in the same struct pinctrl *p.

This fails the sanity check in pinmux.c:

Quoted 13 lines
	/*
	 * If the function selector is already set, it needs to be identical,
	 * we support several groups with one function but not several
	 * functions with one or several groups in the same pinmux.
	 */
	if (p->func_selector != UINT_MAX &&
	    p->func_selector != func_selector) {
		dev_err(pctldev->dev,
			"dual function defines in the map for device %s\n",
		       devname);
		return -EINVAL;
	}
	p->func_selector = func_selector;

Oh right, I'd forgotten about that check at this stage in the patch series.

Later in the series, this restriction is removed; each setting in a state is completely independent, and could easily program a different function onto each group, or the same function. At this stage in the series, a struct pinctrl can't represent everything that a set of mapping table entries can, but later struct pinctrl is fully capable.

I guess the solution here is to just make this patch introduce the PINCTRL_STATE_DEFAULT define, but defer most usage of it until later in the series when this restriction is removed.

Note that if you apply all my patches together, you should find that U300 is unbroken, so you can at least test everything. Actually committing that probably isn't a good idea since it'll break git bisect for a little while though.

Signature · 1 line
nvpublic
#9 lore raw

Stephen Warren wrote at Wednesday, February 29, 2012 10:42 AM:

Quoted 45 lines

Linus Walleij wrote at Wednesday, February 29, 2012 9:41 AM:

Quoted 21 lines

Aha now I see the problem here...

Quoted 1 line · Stephen Warren

This provides a single centralized name for the default state.

(...)

Quoted 2 lines
drivers/pinctrl/core.c +0 -2
-/* Hog a single map entry and add to the hoglist */
-static int pinctrl_hog_map(struct pinctrl_dev *pctldev,

(...)

Quoted 1 line

-static int pinctrl_hog_maps(struct pinctrl_dev *pctldev)

(...)

Quoted 4 lines

-       pinctrl_hog_maps(pctldev);
+       pctldev->p = pinctrl_get(pctldev->dev, PINCTRL_STATE_DEFAULT);
+       if (!IS_ERR(pctldev->p))
+               pinctrl_enable(pctldev->p);

So what happens here is that my hogs will try to activate three different functions in the same struct pinctrl *p.

This fails the sanity check in pinmux.c:

Quoted 13 lines
	/*
	 * If the function selector is already set, it needs to be identical,
	 * we support several groups with one function but not several
	 * functions with one or several groups in the same pinmux.
	 */
	if (p->func_selector != UINT_MAX &&
	    p->func_selector != func_selector) {
		dev_err(pctldev->dev,
			"dual function defines in the map for device %s\n",
		       devname);
		return -EINVAL;
	}
	p->func_selector = func_selector;

Oh right, I'd forgotten about that check at this stage in the patch series.

Later in the series, this restriction is removed; each setting in a state is completely independent, and could easily program a different function onto each group, or the same function. At this stage in the series, a struct pinctrl can't represent everything that a set of mapping table entries can, but later struct pinctrl is fully capable.

Quoted 3 lines

I guess the solution here is to just make this patch introduce the PINCTRL_STATE_DEFAULT define, but defer most usage of it until later in the series when this restriction is removed.

Actually, it looks very easy to move the func_selector field out of struct pinctrl into struct pinmux_group, which would solve this problem in what I consider the correct way. I'll code up a patch to do that and stack it before this current patch.

Signature · 1 line
nvpublic
#10 lore raw
Quoted 4 lines · Stephen Warren

Actually, it looks very easy to move the func_selector field out of struct pinctrl into struct pinmux_group, which would solve this problem in what I consider the correct way. I'll code up a patch to do that and stack it before this current patch.

Hm that sounds really elegant, go for it!

Yours,
Linus Walleij