[PATCH] drivercore: Output common devicetree information in uevent

[PATCH] drivercore: Output common devicetree information in uevent

[Grant Likely]

When userspace needs to find a specific device, it currently isn't easy to
resolve a /sys/devices/ path from a specific device tree node.  Nor is it
easy to obtain the compatible list for devices.

This patch generalizes the code that inserts OF_* values into the uevent
device attribute so that any device that is attached to an OF node will
have that information exported to userspace.  Without this patch only
platform devices and some powerpc-specific busses have access to this
data.

The original function also creates a MODALIAS property for the compatible
list, but that code has not been generalized into the common case because
it has the potential to break module loading on a lot of bus types.  Bus
types are still responsible for their own MODALIAS properties.

Boot tested on ARM and compile tested on PowerPC and SPARC.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Cc: Greg Kroah-Hartman <greg@kroah.com>
Cc: Tobias Klauser <tklauser@distanz.ch>
Cc: Frederic Lambert <frdrc66@gmail.com>
Cc: Rob Herring <rob.herring@calxeda.com>
Cc: Mark Brown <broonie@sirena.org.uk>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---

Resend: actually including the mailing lists this time and using an up-to-date
email address for Greg.

g.

 arch/powerpc/kernel/ibmebus.c  |    2 +-
 drivers/base/core.c            |    5 +++++
 drivers/base/platform.c        |    2 +-
 drivers/macintosh/macio_asic.c |    2 +-
 drivers/of/device.c            |   30 ++++++++++++++++--------------
 include/linux/of_device.h      |    8 ++++++--
 6 files changed, 30 insertions(+), 19 deletions(-)

diff --git a/arch/powerpc/kernel/ibmebus.c b/arch/powerpc/kernel/ibmebus.c
index d39ae60..79bb282 100644
--- a/arch/powerpc/kernel/ibmebus.c
+++ b/arch/powerpc/kernel/ibmebus.c
@@ -713,7 +713,7 @@ static struct dev_pm_ops ibmebus_bus_dev_pm_ops = {
 
 struct bus_type ibmebus_bus_type = {
 	.name      = "ibmebus",
-	.uevent    = of_device_uevent,
+	.uevent    = of_device_uevent_modalias,
 	.bus_attrs = ibmebus_bus_attrs,
 	.match     = ibmebus_bus_bus_match,
 	.probe     = ibmebus_bus_device_probe,
diff --git a/drivers/base/core.c b/drivers/base/core.c
index 4a67cc0..28d8c21 100644
--- a/drivers/base/core.c
+++ b/drivers/base/core.c
@@ -18,6 +18,8 @@
 #include <linux/string.h>
 #include <linux/kdev_t.h>
 #include <linux/notifier.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
 #include <linux/genhd.h>
 #include <linux/kallsyms.h>
 #include <linux/mutex.h>
@@ -267,6 +269,9 @@ static int dev_uevent(struct kset *kset, struct kobject *kobj,
 	if (dev->driver)
 		add_uevent_var(env, "DRIVER=%s", dev->driver->name);
 
+	/* Add common DT information about the device */
+	of_device_uevent(dev, env);
+
 	/* have the bus specific function add its stuff */
 	if (dev->bus && dev->bus->uevent) {
 		retval = dev->bus->uevent(dev, env);
diff --git a/drivers/base/platform.c b/drivers/base/platform.c
index f0c605e..a1a7225 100644
--- a/drivers/base/platform.c
+++ b/drivers/base/platform.c
@@ -621,7 +621,7 @@ static int platform_uevent(struct device *dev, struct kobj_uevent_env *env)
 	int rc;
 
 	/* Some devices have extra OF data and an OF-style MODALIAS */
-	rc = of_device_uevent(dev,env);
+	rc = of_device_uevent_modalias(dev,env);
 	if (rc != -ENODEV)
 		return rc;
 
diff --git a/drivers/macintosh/macio_asic.c b/drivers/macintosh/macio_asic.c
index 4daf9e5..20e5c2c 100644
--- a/drivers/macintosh/macio_asic.c
+++ b/drivers/macintosh/macio_asic.c
@@ -137,7 +137,7 @@ extern struct device_attribute macio_dev_attrs[];
 struct bus_type macio_bus_type = {
        .name	= "macio",
        .match	= macio_bus_match,
-       .uevent = of_device_uevent,
+       .uevent = of_device_uevent_modalias,
        .probe	= macio_device_probe,
        .remove	= macio_device_remove,
        .shutdown = macio_device_shutdown,
diff --git a/drivers/of/device.c b/drivers/of/device.c
index 62b4b32..4c74e4f 100644
--- a/drivers/of/device.c
+++ b/drivers/of/device.c
@@ -128,39 +128,41 @@ ssize_t of_device_get_modalias(struct device *dev, char *str, ssize_t len)
 /**
  * of_device_uevent - Display OF related uevent information
  */
-int of_device_uevent(struct device *dev, struct kobj_uevent_env *env)
+void of_device_uevent(struct device *dev, struct kobj_uevent_env *env)
 {
 	const char *compat;
 	int seen = 0, cplen, sl;
 
 	if ((!dev) || (!dev->of_node))
-		return -ENODEV;
-
-	if (add_uevent_var(env, "OF_NAME=%s", dev->of_node->name))
-		return -ENOMEM;
+		return;
 
-	if (add_uevent_var(env, "OF_TYPE=%s", dev->of_node->type))
-		return -ENOMEM;
+	add_uevent_var(env, "OF_NAME=%s", dev->of_node->name);
+	add_uevent_var(env, "OF_FULLNAME=%s", dev->of_node->full_name);
+	if (dev->of_node->type && strcmp("<NULL>", dev->of_node->type) != 0)
+		add_uevent_var(env, "OF_TYPE=%s", dev->of_node->type);
 
 	/* Since the compatible field can contain pretty much anything
 	 * it's not really legal to split it out with commas. We split it
 	 * up using a number of environment variables instead. */
-
 	compat = of_get_property(dev->of_node, "compatible", &cplen);
 	while (compat && *compat && cplen > 0) {
-		if (add_uevent_var(env, "OF_COMPATIBLE_%d=%s", seen, compat))
-			return -ENOMEM;
-
+		add_uevent_var(env, "OF_COMPATIBLE_%d=%s", seen, compat);
 		sl = strlen(compat) + 1;
 		compat += sl;
 		cplen -= sl;
 		seen++;
 	}
+	add_uevent_var(env, "OF_COMPATIBLE_N=%d", seen);
+}
 
-	if (add_uevent_var(env, "OF_COMPATIBLE_N=%d", seen))
-		return -ENOMEM;
+int of_device_uevent_modalias(struct device *dev, struct kobj_uevent_env *env)
+{
+	int sl;
+
+	if ((!dev) || (!dev->of_node))
+		return -ENODEV;
 
-	/* modalias is trickier, we add it in 2 steps */
+	/* Devicetree modalias is tricky, we add it in 2 steps */
 	if (add_uevent_var(env, "MODALIAS="))
 		return -ENOMEM;
 
diff --git a/include/linux/of_device.h b/include/linux/of_device.h
index ae56384..cbc4214 100644
--- a/include/linux/of_device.h
+++ b/include/linux/of_device.h
@@ -34,7 +34,8 @@ extern void of_device_unregister(struct platform_device *ofdev);
 extern ssize_t of_device_get_modalias(struct device *dev,
 					char *str, ssize_t len);
 
-extern int of_device_uevent(struct device *dev, struct kobj_uevent_env *env);
+extern void of_device_uevent(struct device *dev, struct kobj_uevent_env *env);
+extern int of_device_uevent_modalias(struct device *dev, struct kobj_uevent_env *env);
 
 static inline void of_device_node_put(struct device *dev)
 {
@@ -49,7 +50,10 @@ static inline int of_driver_match_device(struct device *dev,
 	return 0;
 }
 
-static inline int of_device_uevent(struct device *dev,
+static inline void of_device_uevent(struct device *dev,
+			struct kobj_uevent_env *env) { }
+
+static inline int of_device_uevent_modalias(struct device *dev,
 				   struct kobj_uevent_env *env)
 {
 	return -ENODEV;
-- 
1.7.5.4

Re: [PATCH] drivercore: Output common devicetree information in uevent

[Greg Kroah-Hartman]

On Wed, Feb 01, 2012 at 11:39:04AM -0700, Grant Likely wrote:
> When userspace needs to find a specific device, it currently isn't easy to
> resolve a /sys/devices/ path from a specific device tree node.  Nor is it
> easy to obtain the compatible list for devices.
> 
> This patch generalizes the code that inserts OF_* values into the uevent
> device attribute so that any device that is attached to an OF node will
> have that information exported to userspace.  Without this patch only
> platform devices and some powerpc-specific busses have access to this
> data.
> 
> The original function also creates a MODALIAS property for the compatible
> list, but that code has not been generalized into the common case because
> it has the potential to break module loading on a lot of bus types.  Bus
> types are still responsible for their own MODALIAS properties.

So you are only adding new properties to be exported, not removing any
from any existing devices, right?

If so, I have no objection to this at all, if you want to take this
through your tree for 3.4, please feel free to add:

	Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Or I can take it through my driver-core tree, which ever is easiest for
you.

thanks,

greg k-h

Re: [PATCH] drivercore: Output common devicetree information in uevent

[Grant Likely]

On Wed, Feb 1, 2012 at 12:09 PM, Greg Kroah-Hartman <greg@kroah.com> wrote:
> On Wed, Feb 01, 2012 at 11:39:04AM -0700, Grant Likely wrote:
>> When userspace needs to find a specific device, it currently isn't easy to
>> resolve a /sys/devices/ path from a specific device tree node.  Nor is it
>> easy to obtain the compatible list for devices.
>>
>> This patch generalizes the code that inserts OF_* values into the uevent
>> device attribute so that any device that is attached to an OF node will
>> have that information exported to userspace.  Without this patch only
>> platform devices and some powerpc-specific busses have access to this
>> data.
>>
>> The original function also creates a MODALIAS property for the compatible
>> list, but that code has not been generalized into the common case because
>> it has the potential to break module loading on a lot of bus types.  Bus
>> types are still responsible for their own MODALIAS properties.
>
> So you are only adding new properties to be exported, not removing any
> from any existing devices, right?
>
> If so, I have no objection to this at all, if you want to take this
> through your tree for 3.4, please feel free to add:
>
>        Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
>
> Or I can take it through my driver-core tree, which ever is easiest for
> you.

I'll take it via the devicetree branch.  Thanks.

g.

Re: [PATCH] drivercore: Output common devicetree information in uevent

[Grant Likely]

On Wed, Feb 1, 2012 at 2:29 PM, Grant Likely <grant.likely@secretlab.ca> wrote:
> On Wed, Feb 1, 2012 at 2:16 PM, Benjamin Herrenschmidt
> <benh@kernel.crashing.org> wrote:
>> On Wed, 2012-02-01 at 11:31 -0700, Grant Likely wrote:
>>> The original function also creates a MODALIAS property for the
>>> compatible
>>> list, but that code has not been generalized into the common case
>>> because
>>> it has the potential to break module loading on a lot of bus types.
>>> Bus
>>> types are still responsible for their own MODALIAS properties.
>>>
>>> Boot tested on ARM and compile tested on PowerPC and SPARC.
>>>
>>>
>> We also want to generalize adding the devspec property that we have on
>> PCI to all devices no ?
>
> The pci_show_devspec() function?  Yeah, that can be generalized, but
> haven't researched where the best place to hook it in should be.

Is there any point in generalizing that attribute though when the data
is in uevent?

g.

Re: [PATCH] drivercore: Output common devicetree information in uevent

[Grant Likely]

On Wed, Feb 1, 2012 at 2:16 PM, Benjamin Herrenschmidt
<benh@kernel.crashing.org> wrote:
> On Wed, 2012-02-01 at 11:31 -0700, Grant Likely wrote:
>> The original function also creates a MODALIAS property for the
>> compatible
>> list, but that code has not been generalized into the common case
>> because
>> it has the potential to break module loading on a lot of bus types.
>> Bus
>> types are still responsible for their own MODALIAS properties.
>>
>> Boot tested on ARM and compile tested on PowerPC and SPARC.
>>
>>
> We also want to generalize adding the devspec property that we have on
> PCI to all devices no ?

The pci_show_devspec() function?  Yeah, that can be generalized, but
haven't researched where the best place to hook it in should be.

g.

-- 
Grant Likely, B.Sc., P.Eng.
Secret Lab Technologies Ltd.