[PATCH] POWER: jz4740-battery: Protect against concurrent battery readings

[PATCH] POWER: jz4740-battery: Protect against concurrent battery readings

[Lars-Peter Clausen]

We can not handle more then one ADC request at a time to the battery. The patch
adds a mutex around the ADC read code to ensure this.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Cc: stable@kernel.org
---
 drivers/power/jz4740-battery.c |    7 +++++++
 1 files changed, 7 insertions(+), 0 deletions(-)
---
This patch should also go into the 2.6.36.x stable tree

diff --git a/drivers/power/jz4740-battery.c b/drivers/power/jz4740-battery.c
index a8108a7..2bbe04a 100644
--- a/drivers/power/jz4740-battery.c
+++ b/drivers/power/jz4740-battery.c
@@ -47,6 +47,8 @@ struct jz_battery {
 
 	struct power_supply battery;
 	struct delayed_work work;
+
+	struct mutex lock;
 };
 
 static inline struct jz_battery *psy_to_jz_battery(struct power_supply *psy)
@@ -68,6 +70,8 @@ static long jz_battery_read_voltage(struct jz_battery *battery)
 	unsigned long val;
 	long voltage;
 
+	mutex_lock(&battery->lock);
+
 	INIT_COMPLETION(battery->read_completion);
 
 	enable_irq(battery->irq);
@@ -91,6 +95,8 @@ static long jz_battery_read_voltage(struct jz_battery *battery)
 	battery->cell->disable(battery->pdev);
 	disable_irq(battery->irq);
 
+	mutex_unlock(&battery->lock);
+
 	return voltage;
 }
 
@@ -291,6 +297,7 @@ static int __devinit jz_battery_probe(struct platform_device *pdev)
 	jz_battery->pdev = pdev;
 
 	init_completion(&jz_battery->read_completion);
+	mutex_init(&jz_battery->lock);
 
 	INIT_DELAYED_WORK(&jz_battery->work, jz_battery_work);
 
-- 
1.5.6.5

[PATCH] POWER: jz4740-battery: Check if platform_data is supplied

[Lars-Peter Clausen]

Currently platform_data is dereferenced without checking whether it is actually
set, which can lead to kernel crashes.
This patch adds a check which will abort the drivers probe function gracefully
if no platform_data is supplied.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
---
 drivers/power/jz4740-battery.c |    5 +++++
 1 files changed, 5 insertions(+), 0 deletions(-)

diff --git a/drivers/power/jz4740-battery.c b/drivers/power/jz4740-battery.c
index 2bbe04a..5b48215 100644
--- a/drivers/power/jz4740-battery.c
+++ b/drivers/power/jz4740-battery.c
@@ -246,6 +246,11 @@ static int __devinit jz_battery_probe(struct platform_device *pdev)
 	struct jz_battery *jz_battery;
 	struct power_supply *battery;
 
+	if (!pdata) {
+		dev_err(&pdev->dev, "No platform_data supplied\n");
+		return -ENXIO;
+	}
+
 	jz_battery = kzalloc(sizeof(*jz_battery), GFP_KERNEL);
 	if (!jz_battery) {
 		dev_err(&pdev->dev, "Failed to allocate driver structure\n");
-- 
1.5.6.5

Re: [PATCH] POWER: jz4740-battery: Check if platform_data is supplied

[Anton Vorontsov]

On Thu, Nov 11, 2010 at 07:00:53PM +0100, Lars-Peter Clausen wrote:
> Currently platform_data is dereferenced without checking whether it is actually
> set, which can lead to kernel crashes.
> This patch adds a check which will abort the drivers probe function gracefully
> if no platform_data is supplied.
> 
> Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
> ---

Also applied, thanks.

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

Re: [PATCH] POWER: jz4740-battery: Protect against concurrent battery readings

[Anton Vorontsov]

On Thu, Nov 11, 2010 at 07:00:52PM +0100, Lars-Peter Clausen wrote:
> We can not handle more then one ADC request at a time to the battery. The patch
> adds a mutex around the ADC read code to ensure this.
> 
> Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
> Cc: stable@kernel.org
> ---
>  drivers/power/jz4740-battery.c |    7 +++++++
>  1 files changed, 7 insertions(+), 0 deletions(-)
> ---
> This patch should also go into the 2.6.36.x stable tree

Applied, thanks!

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