In case of failure, device_create() returns not NULL but the error code.
The current code checks for non-NULL though which causes kernel oops in
sysfs_create_group() when device_create() fails. Check for error using
IS_ERR() and propagate the error value using PTR_ERR() instead of fixed
-ENODEV code returned now...
Signed-off-by: Sergei Shtylyov <[email protected]>
drivers/gpio/gpiolib.c | 8 ++++----
1 files changed, 4 insertions(+), 4 deletions(-)
Index: linux-2.6/drivers/gpio/gpiolib.c
===================================================================
--- linux-2.6.orig/drivers/gpio/gpiolib.c
+++ linux-2.6/drivers/gpio/gpiolib.c
@@ -661,7 +661,7 @@ int gpio_export(unsigned gpio, bool dire
dev = device_create(&gpio_class, desc->chip->dev, MKDEV(0, 0),
desc, ioname ? ioname : "gpio%d", gpio);
- if (dev) {
+ if (!IS_ERR(dev)) {
if (direction_may_change)
status = sysfs_create_group(&dev->kobj,
&gpio_attr_group);
@@ -679,7 +679,7 @@ int gpio_export(unsigned gpio, bool dire
if (status != 0)
device_unregister(dev);
} else
- status = -ENODEV;
+ status = PTR_ERR(dev);
if (status == 0)
set_bit(FLAG_EXPORT, &desc->flags);
}
@@ -800,11 +800,11 @@ static int gpiochip_export(struct gpio_c
mutex_lock(&sysfs_lock);
dev = device_create(&gpio_class, chip->dev, MKDEV(0, 0), chip,
"gpiochip%d", chip->base);
- if (dev) {
+ if (!IS_ERR(dev)) {
status = sysfs_create_group(&dev->kobj,
&gpiochip_attr_group);
} else
- status = -ENODEV;
+ status = PTR_ERR(dev);
chip->exported = (status == 0);
mutex_unlock(&sysfs_lock);
On Tue, 20 Oct 2009 20:35:23 +0400
Sergei Shtylyov <[email protected]> wrote:
> In case of failure, device_create() returns not NULL but the error code.
> The current code checks for non-NULL though which causes kernel oops in
> sysfs_create_group() when device_create() fails. Check for error using
> IS_ERR() and propagate the error value using PTR_ERR() instead of fixed
> -ENODEV code returned now...
Does anyone notice any missing information here?
/**
* device_create - creates a device and registers it with sysfs
* @class: pointer to the struct class that this device should be registered to
* @parent: pointer to the parent struct device of this new device, if any
* @devt: the dev_t for the char device to be added
* @drvdata: the data to be added to the device for callbacks
* @fmt: string for the device's name
*
* This function can be used by char device classes. A struct device
* will be created in sysfs, registered to the specified class.
*
* A "dev" file will be created, showing the dev_t for the device, if
* the dev_t is not 0,0.
* If a pointer to a parent struct device is passed in, the newly created
* struct device will be a child of that device in sysfs.
* The pointer to the struct device will be returned from the call.
* Any further sysfs files that might be required can be created using this
* pointer.
*
* Note: the struct class passed to this function must have previously
* been created with a call to class_create().
*/
Why do we do this to ourselves?
On Tue, 3 Nov 2009 14:44:01 -0800
Andrew Morton <[email protected]> wrote:
> On Tue, 20 Oct 2009 20:35:23 +0400
> Sergei Shtylyov <[email protected]> wrote:
>
> > In case of failure, device_create() returns not NULL but the error code.
> > The current code checks for non-NULL though which causes kernel oops in
> > sysfs_create_group() when device_create() fails. Check for error using
> > IS_ERR() and propagate the error value using PTR_ERR() instead of fixed
> > -ENODEV code returned now...
>
> Does anyone notice any missing information here?
its what you get for cheap hacks instead of having a ptr_err_t which is a
union so you have to get it right ;)
Alan
On Tue, Nov 03, 2009 at 02:44:01PM -0800, Andrew Morton wrote:
> On Tue, 20 Oct 2009 20:35:23 +0400
> Sergei Shtylyov <[email protected]> wrote:
>
> > In case of failure, device_create() returns not NULL but the error code.
> > The current code checks for non-NULL though which causes kernel oops in
> > sysfs_create_group() when device_create() fails. Check for error using
> > IS_ERR() and propagate the error value using PTR_ERR() instead of fixed
> > -ENODEV code returned now...
>
> Does anyone notice any missing information here?
>
> /**
> * device_create - creates a device and registers it with sysfs
> * @class: pointer to the struct class that this device should be registered to
> * @parent: pointer to the parent struct device of this new device, if any
> * @devt: the dev_t for the char device to be added
> * @drvdata: the data to be added to the device for callbacks
> * @fmt: string for the device's name
> *
> * This function can be used by char device classes. A struct device
> * will be created in sysfs, registered to the specified class.
> *
> * A "dev" file will be created, showing the dev_t for the device, if
> * the dev_t is not 0,0.
> * If a pointer to a parent struct device is passed in, the newly created
> * struct device will be a child of that device in sysfs.
> * The pointer to the struct device will be returned from the call.
> * Any further sysfs files that might be required can be created using this
> * pointer.
> *
> * Note: the struct class passed to this function must have previously
> * been created with a call to class_create().
> */
>
>
> Why do we do this to ourselves?
Because we suck at writing documentation? :)
Patches to add the one line:
"If an error happens, a PTR_ERR is returned instead of a pointer
to the device"
would be appreciated.
And yes, I like Alan's idea, that would be much nicer overall for lots
of functions like this.
thanks,
greg k-h