Hi,
On 12/16/22 14:55, Andy Shevchenko wrote:
> On Fri, Dec 16, 2022 at 03:35:29PM +0200, Andy Shevchenko wrote:
>> On Fri, Dec 16, 2022 at 12:30:03PM +0100, Hans de Goede wrote:
>
> ...
>
>>> led_cdev = dev_get_drvdata(led_dev);
>>>
>>> - if (!try_module_get(led_cdev->dev->parent->driver->owner))
>>> + if (!try_module_get(led_cdev->dev->parent->driver->owner)) {
>>> + put_device(led_cdev->dev);
>>> return ERR_PTR(-ENODEV);
>>> + }
>>>
>>> return led_cdev;
>>
>> ...
>>
>>> void led_put(struct led_classdev *led_cdev)
>>> {
>>> module_put(led_cdev->dev->parent->driver->owner);
>>> + put_device(led_cdev->dev);
>>
>> Hmm... It was in the original submission.
>>
>> https://lore.kernel.org/linux-leds/1443605522-1118-2-git-send-email-tomi.valkeinen@xxxxxx/
>
> ...
>
>> Nevertheless, shouldn't you put device before putting module? (It may need to
>> save the owner of the driver, I think.)
>
> I think this is wrong, the symmetry is kept correct in your patch.
Right, the line above dereferences led_cdev->dev, so the put()
must be done after that line.
Regards,
Hans
