Re: [PATCH 7/7] driver-core : convert semaphore to mutex in struct class

[Jarek Poplawski <jarkao2@gmail.com>]

On Tue, Jan 15, 2008 at 05:15:27PM +0800, Dave Young wrote:
> Convert the class semaphore to mutex.
> 
> Signed-off-by: Dave Young <hidave.darkstar@gmail.com> 
> 
> ---
>  drivers/base/class.c   |   38 +++++++++++++++++++-------------------
>  drivers/base/core.c    |   18 ++++++++----------
>  include/linux/device.h |    3 ++-
>  3 files changed, 29 insertions(+), 30 deletions(-)
> 
> diff -upr linux/drivers/base/class.c linux.new/drivers/base/class.c
> --- linux/drivers/base/class.c	2008-01-15 14:04:26.000000000 +0800
> +++ linux.new/drivers/base/class.c	2008-01-15 14:04:26.000000000 +0800
> @@ -144,7 +144,7 @@ int class_register(struct class * cls)
>  	INIT_LIST_HEAD(&cls->devices);
>  	INIT_LIST_HEAD(&cls->interfaces);
>  	kset_init(&cls->class_dirs);
> -	init_MUTEX(&cls->sem);
> +	mutex_init(&cls->mutex);
>  	error = kobject_set_name(&cls->subsys.kobj, "%s", cls->name);
>  	if (error)
>  		return error;
> @@ -617,13 +617,13 @@ int class_device_add(struct class_device
>  	kobject_uevent(&class_dev->kobj, KOBJ_ADD);
>  
>  	/* notify any interfaces this device is now here */
> -	down(&parent_class->sem);
> +	mutex_lock_nested(&parent_class->mutex, SINGLE_DEPTH_NESTING);
>  	list_add_tail(&class_dev->node, &parent_class->children);
>  	list_for_each_entry(class_intf, &parent_class->interfaces, node) {
>  		if (class_intf->add)
>  			class_intf->add(class_dev, class_intf);
>  	}
> -	up(&parent_class->sem);
> +	mutex_unlock(&parent_class->mutex);
>  
>  	goto out1;
>  
> @@ -725,12 +725,12 @@ void class_device_del(struct class_devic
>  	struct class_interface *class_intf;
>  
>  	if (parent_class) {
> -		down(&parent_class->sem);
> +		mutex_lock(&parent_class->mutex);

I hope I'm wrong with this (I don't know this code at all...), and
of course I should've noticed this earlier after all, but I wonder
about this _NESTING corretness here. So, if these variables names
are right, and say about real nesting dependency, then it seems
mutex_lock_nested() should be used consistently even if (currently?)
not forced by lockdep warnings; otherwise this could possibly cover
some other warnings. Alas, if accidentally I'm right, it seems a
bit of new testing would be necessary...

Regards,
Jarek P.