Re: [RFC PATCH v3] Use kernfs_break_active_protection() for device online store callbacks

[Tejun Heo <tj@kernel.org>]

Hello,

On Mon, Apr 14, 2014 at 03:47:29PM +0800, Li Zhong wrote:
> @@ -439,6 +439,7 @@ static ssize_t online_store(struct device *dev, struct device_attribute *attr,
>  {
>  	bool val;
>  	int ret;
> +	struct kernfs_node *kn;
>  
>  	ret = strtobool(buf, &val);
>  	if (ret < 0)
> @@ -448,7 +449,19 @@ static ssize_t online_store(struct device *dev, struct device_attribute *attr,
>  	if (ret)
>  		return ret;
>  
> +	kn = kernfs_find_and_get(dev->kobj.sd, attr->attr.name);
> +	if (WARN_ON_ONCE(!kn))
> +		goto out;
> +
> +	get_device(dev);
> +	kernfs_break_active_protection(kn);
>  	ret = val ? device_online(dev) : device_offline(dev);
> +	kernfs_unbreak_active_protection(kn);
> +	put_device(dev);
> +
> +	kernfs_put(kn);
> +
> +out:
>  	unlock_device_hotplug();
>  	return ret < 0 ? ret : count;
>  }

Can you please add comment explainin why this is being down?  As it
currently stands, it's quite a bit of complexity without any
indication what's going on why.  Also, if device_hotplug is locked, is
it really necessary to get @dev?  Can it go away inbetween?  The code
snippet looks weird because getting @dev indicates that the device
might go away without doing it but then it proceeds to invoke
device_{on|off}line() which probably isn't safe to invoke on a removed
device.

Thanks.

-- 
tejun