Re: [PATCH] power-supply: Add function to return system-wide power state

Previous thread: [PATCH] char: moxa.c trivial sparse annotation by Harvey Harrison on Tuesday, August 26, 2008 - 1:06 pm. (1 message)

Next thread: __kernel_suseconds_t is int on PA-RISC (and SPARC64)? by Geert Uytterhoeven on Tuesday, August 26, 2008 - 1:11 pm. (5 messages)
From: Matthew Garrett
Date: Tuesday, August 26, 2008 - 1:09 pm

(resend, I screwed up the Cc)

Certain drivers benefit from knowing whether the system is on ac or 
battery, for instance when determining which backlight registers to 
read. This adds a simple call to determine whether there's an online 
power supply other than any batteries.

Signed-off-by: Matthew Garrett <mjg@redhat.com>

---

The name is a bit odd, but seems to fit in with the existing naming 
scheme.

diff --git a/drivers/power/power_supply_core.c b/drivers/power/power_supply_core.c
index cb1ccb4..6a687da 100644
--- a/drivers/power/power_supply_core.c
+++ b/drivers/power/power_supply_core.c
@@ -87,6 +87,31 @@ int power_supply_am_i_supplied(struct power_supply *psy)
 	return error;
 }
 
+static int __power_supply_is_system_supplied(struct device *dev, void *data)
+{
+	union power_supply_propval ret = {0,};
+	struct power_supply *psy = dev_get_drvdata(dev);
+
+	if (psy->type != POWER_SUPPLY_TYPE_BATTERY) {
+		if (psy->get_property(psy,
+				       POWER_SUPPLY_PROP_ONLINE, &ret))
+			return 0;
+		if (ret.intval)
+			return ret.intval;
+	}
+	return 0;
+}
+
+int power_supply_is_system_supplied()
+{
+	int error;
+
+	error = class_for_each_device(power_supply_class, NULL, NULL,
+				      __power_supply_is_system_supplied);
+
+	return error;
+}
+
 int power_supply_register(struct device *parent, struct power_supply *psy)
 {
 	int rc = 0;
@@ -148,6 +173,7 @@ static void __exit power_supply_class_exit(void)
 
 EXPORT_SYMBOL_GPL(power_supply_changed);
 EXPORT_SYMBOL_GPL(power_supply_am_i_supplied);
+EXPORT_SYMBOL_GPL(power_supply_is_system_supplied);
 EXPORT_SYMBOL_GPL(power_supply_register);
 EXPORT_SYMBOL_GPL(power_supply_unregister);
 
diff --git a/include/linux/power_supply.h b/include/linux/power_supply.h
index ea96ead..3e67d0c 100644
--- a/include/linux/power_supply.h
+++ b/include/linux/power_supply.h
@@ -164,6 +164,7 @@ struct power_supply_info {
 
 extern void power_supply_changed(struct power_supply *psy);
 extern int ...
From: Anton Vorontsov
Date: Tuesday, August 26, 2008 - 1:25 pm

I like the patch, much thanks!


May I suggest adding
static int power_supply_is_system_supplied(void) { return -ENOSYS; }
in case of !CONFIG_POWER_SUPPLY? This way drivers may not strictly
depend on power supply class.

And "if (power_supply_is_system_supplied())" will assume that we're
always supplied, which good assumption, I think.

Thanks,

-- 
Anton Vorontsov
email: cbouatmailru@gmail.com
irc://irc.freenode.net/bd2
--

From: Matthew Garrett
Date: Tuesday, August 26, 2008 - 1:33 pm

Something like this?

diff --git a/drivers/power/power_supply_core.c b/drivers/power/power_supply_core.c
index cb1ccb4..6a687da 100644
--- a/drivers/power/power_supply_core.c
+++ b/drivers/power/power_supply_core.c
@@ -87,6 +87,31 @@ int power_supply_am_i_supplied(struct power_supply *psy)
 	return error;
 }
 
+static int __power_supply_is_system_supplied(struct device *dev, void *data)
+{
+	union power_supply_propval ret = {0,};
+	struct power_supply *psy = dev_get_drvdata(dev);
+
+	if (psy->type != POWER_SUPPLY_TYPE_BATTERY) {
+		if (psy->get_property(psy,
+				       POWER_SUPPLY_PROP_ONLINE, &ret))
+			return 0;
+		if (ret.intval)
+			return ret.intval;
+	}
+	return 0;
+}
+
+int power_supply_is_system_supplied()
+{
+	int error;
+
+	error = class_for_each_device(power_supply_class, NULL, NULL,
+				      __power_supply_is_system_supplied);
+
+	return error;
+}
+
 int power_supply_register(struct device *parent, struct power_supply *psy)
 {
 	int rc = 0;
@@ -148,6 +173,7 @@ static void __exit power_supply_class_exit(void)
 
 EXPORT_SYMBOL_GPL(power_supply_changed);
 EXPORT_SYMBOL_GPL(power_supply_am_i_supplied);
+EXPORT_SYMBOL_GPL(power_supply_is_system_supplied);
 EXPORT_SYMBOL_GPL(power_supply_register);
 EXPORT_SYMBOL_GPL(power_supply_unregister);
 
diff --git a/include/linux/power_supply.h b/include/linux/power_supply.h
index ea96ead..304df22 100644
--- a/include/linux/power_supply.h
+++ b/include/linux/power_supply.h
@@ -164,11 +164,16 @@ struct power_supply_info {
 
 extern void power_supply_changed(struct power_supply *psy);
 extern int power_supply_am_i_supplied(struct power_supply *psy);
-
 extern int power_supply_register(struct device *parent,
 				 struct power_supply *psy);
 extern void power_supply_unregister(struct power_supply *psy);
 
+#ifdef CONFIG_POWER_SUPPLY
+extern int power_supply_is_system_supplied(void);
+#else
+static int power_supply_is_system_supplied(void) { return -ENOSYS; }
+#endif
+
 /* For APM emulation, think ...
From: Anton Vorontsov
Date: Tuesday, August 26, 2008 - 2:10 pm

Yup, except that I forgot 'inline' in my example. ;-) I'll add it

-- 
Anton Vorontsov
email: cbouatmailru@gmail.com
irc://irc.freenode.net/bd2
--

From: Anton Vorontsov
Date: Sunday, August 31, 2008 - 4:14 pm

I just applied it to battery-2.6.git with the following changes:


-()
+(void)



Added inline keyword, otherwise gcc might warn about defined but
not used function.


Please test battery-2.6.git tree (or simply next `-next` tree ;-)
if everything is okay.

Thanks again!

-- 
Anton Vorontsov
email: cbouatmailru@gmail.com
irc://irc.freenode.net/bd2
--

Previous thread: [PATCH] char: moxa.c trivial sparse annotation by Harvey Harrison on Tuesday, August 26, 2008 - 1:06 pm. (1 message)

Next thread: __kernel_suseconds_t is int on PA-RISC (and SPARC64)? by Geert Uytterhoeven on Tuesday, August 26, 2008 - 1:11 pm. (5 messages)