Skip to contentSkip to conversation

amba bus devicetree probing

PATCH New Rob Herring 3 participants 19 messages Jun 10, 2011

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

v4 cover amba bus devicetree probing

From: Rob Herring <[email protected]>

This is updated amba bus DT probe support. Amba devices are now handled as a special case in the device tree bus scanning code. Found devices with a "arm,amba-device" compatible property will be created by of_amba_device_create.

The first 2 patches are fixes for of_platform_populate.

Applies to devicetree/next on git://git.secretlab.ca/git/linux-2.6

Compile tested with DT enabled and AMBA bus enabled or disabled.

Rob

Rob Herring (4):
dt: check root nodes for match on of_platform_populate
dt: add EXPORT_SYMBOL to of_platform_populate drivers/amba: create devices from device tree
dt: add amba device creation to platform bus scan

 Documentation/devicetree/bindings/arm/amba.txt |   21 ++++++++++
 drivers/amba/bus.c                             |   51 ++++++++++++++++++++++++
 drivers/of/platform.c                          |    9 ++++
 include/linux/amba/bus.h                       |   18 ++++++++
 4 files changed, 99 insertions(+), 0 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/arm/amba.txt
Signature · 1 line
1.7.4.1
#1 lore raw

PATCH v4 1/4 dt: check root nodes for match on of_platform_populate

From: Rob Herring <[email protected]>

of_platform_populate will create platform devices for the top-level nodes if they have a compatible string, but regardless of a match or not.

---
 drivers/of/platform.c |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)
drivers/of/platform.c +2 -0
index 9b785be..dc56a77 100644
@@ -318,6 +318,8 @@ int of_platform_populate(struct device_node *root,
 		return -EINVAL;
 
 	for_each_child_of_node(root, child) {
+		if (!of_match_node(matches, child))
+			continue;
 		rc = of_platform_bus_create(child, matches, parent, true);
 		if (rc)
 			break;
Signature · 1 line
1.7.4.1
#6 lore raw

PATCH v4 2/4 dt: add EXPORT_SYMBOL to of_platform_populate

In reply to Rob Herring

From: Rob Herring <[email protected]>

of_platform_populate is missing an EXPORT_SYMBOL, so add it.

---
 drivers/of/platform.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)
drivers/of/platform.c +1 -0
index dc56a77..8d0f477 100644
@@ -328,4 +328,5 @@ int of_platform_populate(struct device_node *root,
 	of_node_put(root);
 	return rc;
 }
+EXPORT_SYMBOL(of_platform_populate);
 #endif /* !CONFIG_SPARC */
Signature · 1 line
1.7.4.1
#10 lore raw

PATCH v4 3/4 drivers/amba: create devices from device tree

In reply to Rob Herring

From: Rob Herring <[email protected]>

Add a function to create amba bus devices (i.e. primecell peripherals) from device tree nodes. The device tree scanning is done by of_platform_probe/populate functions which can call of_amba_device_create based on a match table entry.

Nodes with a "arm,amba-deviceid" property can override the h/w peripheral id value.

Based on the original work by Jeremy Kerr.

Cc Jeremy Kerr <[email protected]>
Cc Grant Likely <[email protected]>
Cc linux at arm.linux.org.uk
Cc arnd at arndb.de
---
 Documentation/devicetree/bindings/arm/amba.txt |   21 ++++++++++
 drivers/amba/bus.c                             |   51 ++++++++++++++++++++++++
 include/linux/amba/bus.h                       |   18 ++++++++
 3 files changed, 90 insertions(+), 0 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/arm/amba.txt
Documentation/devicetree/bindings/arm/amba.txt new file +21 -0
new file mode 100644
index 0000000..23fde7f
@@ -0,0 +1,21 @@
+* ARM Primecell Peripherals
+
+ARM, Ltd. Primecell peripherals have a standard id register that can be used to
+identify the peripheral type, vendor, and revision. This value can be used for
+driver matching. 
+
+Required properties:
+
+- compatible : should be a specific value for peripheral and "arm,amba-device"
+
+Optional properties:
+
+- arm,amba-deviceid : Value to override the h/w value with
+
+Example:
+
+serial at fff36000 {
+	compatible = "arm,pl011", "arm,amba-device";
+	arm,amba-deviceid = <0x00341011>;
+};
+
drivers/amba/bus.c +51 -0

60 lines not on this page · raw message

include/linux/amba/bus.h +18 -0
index fcbbe71..0177274 100644
@@ -24,6 +24,7 @@
 #define AMBA_CID	0xb105f00d
 
 struct clk;
+struct device_node;
 
 struct amba_device {
 	struct device		dev;
@@ -63,6 +64,7 @@ extern struct bus_type amba_bustype;
 #define amba_get_drvdata(d)	dev_get_drvdata(&d->dev)
 #define amba_set_drvdata(d,p)	dev_set_drvdata(&d->dev, p)
 
+#ifdef CONFIG_ARM_AMBA
 int amba_driver_register(struct amba_driver *);
 void amba_driver_unregister(struct amba_driver *);
 int amba_device_register(struct amba_device *, struct resource *);
@@ -94,4 +96,20 @@ void amba_release_regions(struct amba_device *);
 #define amba_manf(d)	AMBA_MANF_BITS((d)->periphid)
 #define amba_part(d)	AMBA_PART_BITS((d)->periphid)
 
+#ifdef CONFIG_OF
+int of_amba_device_create(struct device_node *node, struct device *parent);
+#endif
+
+#else
+
+#ifdef CONFIG_OF
+static inline int of_amba_device_create(struct device_node *node, 
+					struct device *parent)
+{
+	return 0;
+}
+#endif
+
+#endif /* CONFIG_ARM_AMBA */
+
 #endif
Signature · 1 line
1.7.4.1
#16 lore raw

PATCH v4 4/4 dt: add amba device creation to platform bus scan

In reply to Rob Herring

From: Rob Herring <[email protected]>

Add support to the platform bus scanning to call custom device creation function for amba devices.

Cc Jeremy Kerr <[email protected]>
Cc Grant Likely <[email protected]>
Cc arnd at arndb.de
---
 drivers/of/platform.c |    6 ++++++
 1 files changed, 6 insertions(+), 0 deletions(-)
drivers/of/platform.c +6 -0
index 8d0f477..1712f22 100644
@@ -21,6 +21,7 @@
 #include <linux/of_irq.h>
 #include <linux/of_platform.h>
 #include <linux/platform_device.h>
+#include <linux/amba/bus.h>
 
 static int of_dev_node_match(struct device *dev, void *data)
 {
@@ -234,6 +235,11 @@ static int of_platform_bus_create(struct device_node *bus,
 		return 0;
 	}
 
+	if (of_device_is_compatible(bus, "arm,amba-device")) {
+		of_amba_device_create(bus, parent);
+		return 0;
+	}
+	
 	dev = of_platform_device_create(bus, NULL, parent);
 	if (!dev || !of_match_node(matches, bus))
 		return 0;
Signature · 1 line
1.7.4.1
#17 lore raw

PATCH v4 4/4 dt: add amba device creation to platform bus scan

Quoted 10 lines · On Friday 10 June 2011 22:48:43 Rob Herring wrote:

From: Rob Herring <[email protected]>

Add support to the platform bus scanning to call custom device creation function for amba devices.

Cc Jeremy Kerr <[email protected]>
Cc Grant Likely <[email protected]>
Cc arnd at arndb.de

Looks so trivial now ;-)

#11 lore raw

PATCH v4 3/4 drivers/amba: create devices from device tree

In reply to Rob Herring
Quoted 5 lines · On Friday 10 June 2011 22:48:42 Rob Herring wrote:
include/linux/amba/bus.h +4 -0
+#ifdef CONFIG_OF
+int of_amba_device_create(struct device_node *node, struct device *parent);
+#endif
+

include/linux/amba/bus.h void amba_release_regions(struct amba_device *); · line 112

Quoted 10 lines
include/linux/amba/bus.h +10 -0
+#else
+
+#ifdef CONFIG_OF
+static inline int of_amba_device_create(struct device_node *node, 
+                                       struct device *parent)
+{
+       return 0;
+}
+#endif
+

We normally don't use #ifdef around declarations like this, it doesn't hurt to have them even when the code is not there. Otherwise the patch looks good.

Arnd

#2 lore raw

PATCH v4 1/4 dt: check root nodes for match on of_platform_populate

In reply to Rob Herring

drivers/of/platform.c int of_platform_populate(struct device_node *root, · line 325

Quoted 13 lines · On Friday 10 June 2011 22:48:40 Rob Herring wrote:
drivers/of/platform.c +2 -0
index 9b785be..dc56a77 100644
@@ -318,6 +318,8 @@ int of_platform_populate(struct device_node *root,
                return -EINVAL;
 
        for_each_child_of_node(root, child) {
+               if (!of_match_node(matches, child))
+                       continue;
                rc = of_platform_bus_create(child, matches, parent, true);
                if (rc)
                        break;

Why doesn't of_platform_populate just call of_platform_bus_create on the root node? That should normally take care of matching all children.

Arnd

#3 lore raw

PATCH v4 1/4 dt: check root nodes for match on of_platform_populate

Quoted 8 lines · Arnd Bergmann
Quoted 7 lines · On Friday 10 June 2011 22:48:40 Rob Herring wrote:
drivers/of/platform.c +0 -0
index 9b785be..dc56a77 100644
@@ -318,6 +318,8 @@ int of_platform_populate(struct device_node *root,
                return -EINVAL;
 
Quoted 10 lines

drivers/of/platform.c int of_platform_populate(struct device_node *root, · line 325

Quoted 6 lines
drivers/of/platform.c +2 -0
        for_each_child_of_node(root, child) {
+               if (!of_match_node(matches, child))
+                       continue;
                rc = of_platform_bus_create(child, matches, parent, true);
                if (rc)
                        break;

Why doesn't of_platform_populate just call of_platform_bus_create on the root node? That should normally take care of matching all children.

of_platform_bus_create creates a device for the 'bus' node, where of_platform_populate() creates children as direct children of the 'parent' device; which I think is the behaviour wanted here.

Also, of_platform_bus_create checks for a compatible property on the bus node, whereas of_platform_populate does not, and should not.

g.

#5 lore raw

PATCH v4 1/4 dt: check root nodes for match on of_platform_populate

In reply to Rob Herring
Quoted 16 lines · Rob Herring

From: Rob Herring <[email protected]>

of_platform_populate will create platform devices for the top-level nodes if they have a compatible string, but regardless of a match or not.

---
 drivers/of/platform.c |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)
drivers/of/platform.c +0 -0
index 9b785be..dc56a77 100644

drivers/of/platform.c int of_platform_populate(struct device_node *root, · line 322

Quoted 5 lines
drivers/of/platform.c +2 -0
@@ -318,6 +318,8 @@ int of_platform_populate(struct device_node *root,
 		return -EINVAL;
 
 	for_each_child_of_node(root, child) {
+		if (!of_match_node(matches, child))
+			continue;

This doesn't look right. The point of of_platform_populate() is to register platform_devices for all the child nodes, regardless of whether or not they are in the match table.

of_platform_bus_create() does the correct behaviour to check for whether or not to register grandchildren.

Quoted 6 lines
drivers/of/platform.c +0 -0
 		rc = of_platform_bus_create(child, matches, parent, true);
 		if (rc)
 			break;
Signature · 1 line
1.7.4.1
#4 lore raw

PATCH v4 1/4 dt: check root nodes for match on of_platform_populate

In reply to Grant Likely
Quoted 9 lines · On Friday 10 June 2011 23:11:19 Grant Likely wrote:
Quoted 2 lines

Why doesn't of_platform_populate just call of_platform_bus_create on the root node? That should normally take care of matching all children.

of_platform_bus_create creates a device for the 'bus' node, where of_platform_populate() creates children as direct children of the 'parent' device; which I think is the behaviour wanted here.

Also, of_platform_bus_create checks for a compatible property on the bus node, whereas of_platform_populate does not, and should not.

Ok, makes sense.

Thanks,

Arnd

#7 lore raw

PATCH v4 2/4 dt: add EXPORT_SYMBOL to of_platform_populate

In reply to Rob Herring
Quoted 15 lines · Rob Herring

From: Rob Herring <[email protected]>

of_platform_populate is missing an EXPORT_SYMBOL, so add it.

---
 drivers/of/platform.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)
drivers/of/platform.c +0 -0
index dc56a77..8d0f477 100644

drivers/of/platform.c int of_platform_populate(struct device_node *root, · line 331

Quoted 4 lines
drivers/of/platform.c +1 -0
@@ -328,4 +328,5 @@ int of_platform_populate(struct device_node *root,
 	of_node_put(root);
 	return rc;
 }
+EXPORT_SYMBOL(of_platform_populate);

I hadn't done this because I hadn't seen a need for calling it from module code. Do you have a driver that uses it?

g.

#13 lore raw

PATCH v4 3/4 drivers/amba: create devices from device tree

In reply to Rob Herring
Quoted 31 lines · Rob Herring

From: Rob Herring <[email protected]>

Add a function to create amba bus devices (i.e. primecell peripherals) from device tree nodes. The device tree scanning is done by of_platform_probe/populate functions which can call of_amba_device_create based on a match table entry.

Nodes with a "arm,amba-deviceid" property can override the h/w peripheral id value.

Based on the original work by Jeremy Kerr.

Cc Jeremy Kerr <[email protected]>
Cc Grant Likely <[email protected]>
Cc linux at arm.linux.org.uk
Cc arnd at arndb.de
---
 Documentation/devicetree/bindings/arm/amba.txt |   21 ++++++++++
 drivers/amba/bus.c                             |   51 ++++++++++++++++++++++++
 include/linux/amba/bus.h                       |   18 ++++++++
 3 files changed, 90 insertions(+), 0 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/arm/amba.txt
Documentation/devicetree/bindings/arm/amba.txt new file +0 -0
new file mode 100644
index 0000000..23fde7f

Documentation/devicetree/bindings/arm/amba.txt line 9

Quoted 9 lines
Documentation/devicetree/bindings/arm/amba.txt new file +9 -0
@@ -0,0 +1,21 @@
+* ARM Primecell Peripherals
+
+ARM, Ltd. Primecell peripherals have a standard id register that can be used to
+identify the peripheral type, vendor, and revision. This value can be used for
+driver matching. 
+
+Required properties:
+
+- compatible : should be a specific value for peripheral and "arm,amba-device"

After recent conversations, I'd like to change this to "arm,primecell" to better reflect the kind of devices it models.

Documentation/devicetree/bindings/arm/amba.txt line 13

Quoted 4 lines
Documentation/devicetree/bindings/arm/amba.txt new file +4 -0
+
+Optional properties:
+
+- arm,amba-deviceid : Value to override the h/w value with

Similarly, based on the primecell device documents, this should probabaly be arm,primecell-periphid.

Quoted 63 lines
Documentation/devicetree/bindings/arm/amba.txt new file +8 -0
+
+Example:
+
+serial at fff36000 {
+	compatible = "arm,pl011", "arm,amba-device";
+	arm,amba-deviceid = <0x00341011>;
+};
+
drivers/amba/bus.c +40 -0
index d74926e..19f712b 100644
@@ -13,6 +13,11 @@
 #include <linux/string.h>
 #include <linux/slab.h>
 #include <linux/io.h>
+#include <linux/of.h>
+#include <linux/of_irq.h>
+#include <linux/of_address.h>
+#include <linux/of_device.h>
+#include <linux/of_platform.h>
 #include <linux/pm.h>
 #include <linux/pm_runtime.h>
 #include <linux/amba/bus.h>
@@ -785,3 +790,49 @@ EXPORT_SYMBOL(amba_device_unregister);
 EXPORT_SYMBOL(amba_find_device);
 EXPORT_SYMBOL(amba_request_regions);
 EXPORT_SYMBOL(amba_release_regions);
+
+#ifdef CONFIG_OF
+int of_amba_device_create(struct device_node *node, struct device *parent)
+{
+	struct amba_device *dev;
+	const void *prop;
+	int i, ret;
+
+	dev = kzalloc(sizeof(*dev), GFP_KERNEL);
+	if (!dev)
+		return -ENOMEM;
+
+	/* setup generic device info */
+	dev->dev.coherent_dma_mask = ~0;
+	dev->dev.of_node = node;
+	dev->dev.parent = parent;
+	of_device_make_bus_id(&dev->dev);
+
+	/* setup amba-specific device info */
+	dev->dma_mask = ~0;
+
+	/* Allow the arm,amba-deviceid value to override the h/w value */
+	prop = of_get_property(node, "arm,amba-deviceid", NULL);
+	if (prop)
+		dev->periphid = of_read_ulong(prop, 1);
+
+	/* Decode the IRQs and address ranges */
+	for (i = 0; i < AMBA_NR_IRQS; i++)
+		dev->irq[i] = irq_of_parse_and_map(node, i);
+
+	ret = of_address_to_resource(node, 0, &dev->res);
+	if (ret)
+		goto err_free;
+
+	ret = amba_device_register(dev, &iomem_resource);

drivers/amba/bus.c EXPORT_SYMBOL(amba_device_unregister); · line 837

Quoted 10 lines
drivers/amba/bus.c +10 -0
+	if (ret)
+		goto err_free;
+
+	return 0;
+
+err_free:
+	kfree(dev);
+	return ret;
+}
+

Yeah, I think this looks like the right thing to do. I'll pick it up into devicetree/test and probably move it to devicetree/arm in the near future after testing a bit.

Quoted 45 lines
drivers/amba/bus.c +1 -0
+#endif /* CONFIG_OF */
include/linux/amba/bus.h +18 -0
index fcbbe71..0177274 100644
@@ -24,6 +24,7 @@
 #define AMBA_CID	0xb105f00d
 
 struct clk;
+struct device_node;
 
 struct amba_device {
 	struct device		dev;
@@ -63,6 +64,7 @@ extern struct bus_type amba_bustype;
 #define amba_get_drvdata(d)	dev_get_drvdata(&d->dev)
 #define amba_set_drvdata(d,p)	dev_set_drvdata(&d->dev, p)
 
+#ifdef CONFIG_ARM_AMBA
 int amba_driver_register(struct amba_driver *);
 void amba_driver_unregister(struct amba_driver *);
 int amba_device_register(struct amba_device *, struct resource *);
@@ -94,4 +96,20 @@ void amba_release_regions(struct amba_device *);
 #define amba_manf(d)	AMBA_MANF_BITS((d)->periphid)
 #define amba_part(d)	AMBA_PART_BITS((d)->periphid)
 
+#ifdef CONFIG_OF
+int of_amba_device_create(struct device_node *node, struct device *parent);
+#endif
+
+#else
+
+#ifdef CONFIG_OF
+static inline int of_amba_device_create(struct device_node *node, 
+					struct device *parent)
+{
+	return 0;
+}
+#endif
+
+#endif /* CONFIG_ARM_AMBA */
+
 #endif
Signature · 1 line
1.7.4.1
#12 lore raw

PATCH v4 3/4 drivers/amba: create devices from device tree

In reply to Arnd Bergmann
Quoted 4 lines · Arnd Bergmann

include/linux/amba/bus.h void amba_release_regions(struct amba_device *); · line 100

Quoted 3 lines · On Friday 10 June 2011 22:48:42 Rob Herring wrote:
include/linux/amba/bus.h +2 -0
+#ifdef CONFIG_OF
+int of_amba_device_create(struct device_node *node, struct device *parent);

Actually, it would probably be good to have of_amba_device_create() return a pointer to the created amba_device.

Quoted 8 lines
Quoted 8 lines
include/linux/amba/bus.h +8 -0
+#endif
+
+#else
+
+#ifdef CONFIG_OF
+static inline int of_amba_device_create(struct device_node *node, 
+                                       struct device *parent)
+{
Quoted 8 lines

include/linux/amba/bus.h void amba_release_regions(struct amba_device *); · line 112

Quoted 4 lines
include/linux/amba/bus.h +4 -0
+       return 0;
+}
+#endif
+

We normally don't use #ifdef around declarations like this, it doesn't hurt to have them even when the code is not there. Otherwise the patch looks good.

Yes, this hook will never get called or referenced by anything if CONFIG_OF is not selected. The only real user of it will be drivers/of/platform.c

g.

#14 lore raw

PATCH v4 3/4 drivers/amba: create devices from device tree

In reply to Grant Likely
Quoted 36 lines · Grant Likely
Quoted 35 lines · Rob Herring

From: Rob Herring<[email protected]>

Add a function to create amba bus devices (i.e. primecell peripherals) from device tree nodes. The device tree scanning is done by of_platform_probe/populate functions which can call of_amba_device_create based on a match table entry.

Nodes with a "arm,amba-deviceid" property can override the h/w peripheral id value.

Based on the original work by Jeremy Kerr.

Cc Jeremy Kerr<[email protected]>
Cc Grant Likely<[email protected]>
Cc linux at arm.linux.org.uk
Cc arnd at arndb.de
---
  Documentation/devicetree/bindings/arm/amba.txt |   21 ++++++++++
  drivers/amba/bus.c                             |   51 ++++++++++++++++++++++++
  include/linux/amba/bus.h                       |   18 ++++++++
  3 files changed, 90 insertions(+), 0 deletions(-)

create mode 100644 Documentation/devicetree/bindings/arm/amba.txt

Documentation/devicetree/bindings/arm/amba.txt new file +4 -0
new file mode 100644
index 0000000..23fde7f
@@ -0,0 +1,21 @@
+* ARM Primecell Peripherals
+
+ARM, Ltd. Primecell peripherals have a standard id register that can be used to
+identify the peripheral type, vendor, and revision. This value can be used for
Quoted 8 lines

Documentation/devicetree/bindings/arm/amba.txt line 9

Quoted 5 lines
Documentation/devicetree/bindings/arm/amba.txt new file +5 -0
+driver matching.
+
+Required properties:
+
+- compatible : should be a specific value for peripheral and "arm,amba-device"

After recent conversations, I'd like to change this to "arm,primecell" to better reflect the kind of devices it models.

That did cross my mind as I was writing the doc...

Quoted 78 lines

Documentation/devicetree/bindings/arm/amba.txt line 13

Quoted 5 lines
Documentation/devicetree/bindings/arm/amba.txt new file +4 -0
+
+Optional properties:
+
+- arm,amba-deviceid : Value to override the h/w value with

Similarly, based on the primecell device documents, this should probabaly be arm,primecell-periphid.

Quoted 69 lines
Documentation/devicetree/bindings/arm/amba.txt new file +8 -0
+
+Example:
+
+serial at fff36000 {
+	compatible = "arm,pl011", "arm,amba-device";
+	arm,amba-deviceid =<0x00341011>;
+};
+
drivers/amba/bus.c +46 -0
index d74926e..19f712b 100644
@@ -13,6 +13,11 @@
  #include<linux/string.h>
  #include<linux/slab.h>
  #include<linux/io.h>
+#include<linux/of.h>
+#include<linux/of_irq.h>
+#include<linux/of_address.h>
+#include<linux/of_device.h>
+#include<linux/of_platform.h>
  #include<linux/pm.h>
  #include<linux/pm_runtime.h>
  #include<linux/amba/bus.h>
@@ -785,3 +790,49 @@ EXPORT_SYMBOL(amba_device_unregister);
  EXPORT_SYMBOL(amba_find_device);
  EXPORT_SYMBOL(amba_request_regions);
  EXPORT_SYMBOL(amba_release_regions);
+
+#ifdef CONFIG_OF
+int of_amba_device_create(struct device_node *node, struct device *parent)
+{
+	struct amba_device *dev;
+	const void *prop;
+	int i, ret;
+
+	dev = kzalloc(sizeof(*dev), GFP_KERNEL);
+	if (!dev)
+		return -ENOMEM;
+
+	/* setup generic device info */
+	dev->dev.coherent_dma_mask = ~0;
+	dev->dev.of_node = node;
+	dev->dev.parent = parent;
+	of_device_make_bus_id(&dev->dev);
+
+	/* setup amba-specific device info */
+	dev->dma_mask = ~0;
+
+	/* Allow the arm,amba-deviceid value to override the h/w value */
+	prop = of_get_property(node, "arm,amba-deviceid", NULL);
+	if (prop)
+		dev->periphid = of_read_ulong(prop, 1);
+
+	/* Decode the IRQs and address ranges */
+	for (i = 0; i<  AMBA_NR_IRQS; i++)
+		dev->irq[i] = irq_of_parse_and_map(node, i);
+
+	ret = of_address_to_resource(node, 0,&dev->res);
+	if (ret)
+		goto err_free;
+
+	ret = amba_device_register(dev,&iomem_resource);
+	if (ret)
+		goto err_free;
+
+	return 0;
+
+err_free:
Quoted 8 lines

drivers/amba/bus.c EXPORT_SYMBOL(amba_device_unregister); · line 837

Quoted 4 lines
drivers/amba/bus.c +4 -0
+	kfree(dev);
+	return ret;
+}
+

Yeah, I think this looks like the right thing to do. I'll pick it up into devicetree/test and probably move it to devicetree/arm in the near future after testing a bit.

Okay, I'll respin with your and Arnd's comments.

Rob

#18 lore raw

PATCH v4 4/4 dt: add amba device creation to platform bus scan

In reply to Rob Herring
Quoted 26 lines · Rob Herring

From: Rob Herring <[email protected]>

Add support to the platform bus scanning to call custom device creation function for amba devices.

Cc Jeremy Kerr <[email protected]>
Cc Grant Likely <[email protected]>
Cc arnd at arndb.de
---
 drivers/of/platform.c |    6 ++++++
 1 files changed, 6 insertions(+), 0 deletions(-)
drivers/of/platform.c +1 -0
index 8d0f477..1712f22 100644
@@ -21,6 +21,7 @@
 #include <linux/of_irq.h>
 #include <linux/of_platform.h>
 #include <linux/platform_device.h>
+#include <linux/amba/bus.h>
 
 static int of_dev_node_match(struct device *dev, void *data)
 {

drivers/of/platform.c static int of_platform_bus_create(struct device_node *bus, · line 242

Quoted 8 lines
drivers/of/platform.c +5 -0
@@ -234,6 +235,11 @@ static int of_platform_bus_create(struct device_node *bus,
 		return 0;
 	}
 
+	if (of_device_is_compatible(bus, "arm,amba-device")) {
+		of_amba_device_create(bus, parent);
+		return 0;
+	}
+	

Yeah, this is the right thing to do. I'll pick it up.

Quoted 6 lines
drivers/of/platform.c +0 -0
 	dev = of_platform_device_create(bus, NULL, parent);
 	if (!dev || !of_match_node(matches, bus))
 		return 0;
Signature · 1 line
1.7.4.1
#8 lore raw

PATCH v4 2/4 dt: add EXPORT_SYMBOL to of_platform_populate

In reply to Grant Likely
Quoted 11 lines · Grant Likely
Quoted 10 lines · Rob Herring

From: Rob Herring<[email protected]>

of_platform_populate is missing an EXPORT_SYMBOL, so add it.

---
  drivers/of/platform.c |    1 +
  1 files changed, 1 insertions(+), 0 deletions(-)
Quoted 13 lines

drivers/of/platform.c int of_platform_populate(struct device_node *root, · line 331

Quoted 9 lines
drivers/of/platform.c +1 -0
index dc56a77..8d0f477 100644
@@ -328,4 +328,5 @@ int of_platform_populate(struct device_node *root,
  	of_node_put(root);
  	return rc;
  }
+EXPORT_SYMBOL(of_platform_populate);

I hadn't done this because I hadn't seen a need for calling it from module code. Do you have a driver that uses it?

No, but then it doesn't appear that other functions in this file are used by modules either. Only of_platform_device_create is used by a module (drivers/macintosh) AFAICT.

Either way, it doesn't really matter to me.

Rob

#15 lore raw

PATCH v4 3/4 drivers/amba: create devices from device tree

In reply to Rob Herring
Quoted 135 lines · Rob Herring
Quoted 36 lines · Grant Likely
Quoted 35 lines · Rob Herring

From: Rob Herring<[email protected]>

Add a function to create amba bus devices (i.e. primecell peripherals) from device tree nodes. The device tree scanning is done by of_platform_probe/populate functions which can call of_amba_device_create based on a match table entry.

Nodes with a "arm,amba-deviceid" property can override the h/w peripheral id value.

Based on the original work by Jeremy Kerr.

Cc Jeremy Kerr<[email protected]>
Cc Grant Likely<[email protected]>
Cc linux at arm.linux.org.uk
Cc arnd at arndb.de
---
 Documentation/devicetree/bindings/arm/amba.txt |   21 ++++++++++
 drivers/amba/bus.c                             |   51 ++++++++++++++++++++++++
 include/linux/amba/bus.h                       |   18 ++++++++
 3 files changed, 90 insertions(+), 0 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/arm/amba.txt
Documentation/devicetree/bindings/arm/amba.txt new file +4 -0
new file mode 100644
index 0000000..23fde7f
@@ -0,0 +1,21 @@
+* ARM Primecell Peripherals
+
+ARM, Ltd. Primecell peripherals have a standard id register that can be used to
+identify the peripheral type, vendor, and revision. This value can be used for
Quoted 8 lines

Documentation/devicetree/bindings/arm/amba.txt line 9

Quoted 5 lines
Documentation/devicetree/bindings/arm/amba.txt new file +5 -0
+driver matching.
+
+Required properties:
+
+- compatible : should be a specific value for peripheral and "arm,amba-device"

After recent conversations, I'd like to change this to "arm,primecell" to better reflect the kind of devices it models.

That did cross my mind as I was writing the doc...

Quoted 78 lines

Documentation/devicetree/bindings/arm/amba.txt line 13

Quoted 5 lines
Documentation/devicetree/bindings/arm/amba.txt new file +4 -0
+
+Optional properties:
+
+- arm,amba-deviceid : Value to override the h/w value with

Similarly, based on the primecell device documents, this should probabaly be arm,primecell-periphid.

Quoted 69 lines
Documentation/devicetree/bindings/arm/amba.txt new file +8 -0
+
+Example:
+
+serial at fff36000 {
+	compatible = "arm,pl011", "arm,amba-device";
+	arm,amba-deviceid =<0x00341011>;
+};
+
drivers/amba/bus.c +46 -0
index d74926e..19f712b 100644
@@ -13,6 +13,11 @@
 #include<linux/string.h>
 #include<linux/slab.h>
 #include<linux/io.h>
+#include<linux/of.h>
+#include<linux/of_irq.h>
+#include<linux/of_address.h>
+#include<linux/of_device.h>
+#include<linux/of_platform.h>
 #include<linux/pm.h>
 #include<linux/pm_runtime.h>
 #include<linux/amba/bus.h>
@@ -785,3 +790,49 @@ EXPORT_SYMBOL(amba_device_unregister);
 EXPORT_SYMBOL(amba_find_device);
 EXPORT_SYMBOL(amba_request_regions);
 EXPORT_SYMBOL(amba_release_regions);
+
+#ifdef CONFIG_OF
+int of_amba_device_create(struct device_node *node, struct device *parent)
+{
+	struct amba_device *dev;
+	const void *prop;
+	int i, ret;
+
+	dev = kzalloc(sizeof(*dev), GFP_KERNEL);
+	if (!dev)
+		return -ENOMEM;
+
+	/* setup generic device info */
+	dev->dev.coherent_dma_mask = ~0;
+	dev->dev.of_node = node;
+	dev->dev.parent = parent;
+	of_device_make_bus_id(&dev->dev);
+
+	/* setup amba-specific device info */
+	dev->dma_mask = ~0;
+
+	/* Allow the arm,amba-deviceid value to override the h/w value */
+	prop = of_get_property(node, "arm,amba-deviceid", NULL);
+	if (prop)
+		dev->periphid = of_read_ulong(prop, 1);
+
+	/* Decode the IRQs and address ranges */
+	for (i = 0; i<  AMBA_NR_IRQS; i++)
+		dev->irq[i] = irq_of_parse_and_map(node, i);
+
+	ret = of_address_to_resource(node, 0,&dev->res);
+	if (ret)
+		goto err_free;
+
+	ret = amba_device_register(dev,&iomem_resource);
+	if (ret)
+		goto err_free;
+
+	return 0;
+
+err_free:
Quoted 8 lines

drivers/amba/bus.c EXPORT_SYMBOL(amba_device_unregister); · line 837

Quoted 4 lines
drivers/amba/bus.c +4 -0
+	kfree(dev);
+	return ret;
+}
+

Yeah, I think this looks like the right thing to do. I'll pick it up into devicetree/test and probably move it to devicetree/arm in the near future after testing a bit.

Quoted 1 line

Okay, I'll respin with your and Arnd's comments.

Actually, hold off. I'm working on it right now so I can get it into my tree. I'll take care of the respin.

g.

#9 lore raw

PATCH v4 2/4 dt: add EXPORT_SYMBOL to of_platform_populate

In reply to Rob Herring
Quoted 26 lines · Rob Herring
Quoted 11 lines · Grant Likely
Quoted 10 lines · Rob Herring

From: Rob Herring<[email protected]>

of_platform_populate is missing an EXPORT_SYMBOL, so add it.

---
 drivers/of/platform.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)
Quoted 13 lines

drivers/of/platform.c int of_platform_populate(struct device_node *root, · line 331

Quoted 9 lines
drivers/of/platform.c +1 -0
index dc56a77..8d0f477 100644
@@ -328,4 +328,5 @@ int of_platform_populate(struct device_node *root,
 	of_node_put(root);
 	return rc;
 }
+EXPORT_SYMBOL(of_platform_populate);

I hadn't done this because I hadn't seen a need for calling it from module code. Do you have a driver that uses it?

Quoted 5 lines

No, but then it doesn't appear that other functions in this file are used by modules either. Only of_platform_device_create is used by a module (drivers/macintosh) AFAICT.

Either way, it doesn't really matter to me.

Okay, let's leave this out for the time being. It is a trivial patch to add if someone needs it.

g.