Re: [PATCH v9 1/5] configfs: Allow dynamic group creation

[Daniel Baluta <daniel.baluta@intel.com>]

Christoph, one question. See below.

On Fri, Oct 23, 2015 at 3:48 PM, Christoph Hellwig <hch@lst.de> wrote:
> Hi Daniel,
>
> all these functions look great, but can you also move your new
> iio_sw_trigger_type_configfs_register/unregister functions to the
> core code as those are the wrappers that everyone would have to write,
> e.g. something like:
>
> int configfs_register_default_group(struct config_group *parent_group,
>                 const char *name, struct config_item_type *item_type)
> {
>         struct config_group *group;
>
>         group = kzalloc(sizeof(*group), GFP_KERNEL);
>         if (!group)
>                 return -ENOMEM;
>         config_group_init_type_name(group, name, item_type);

Until this point this is the implementation of configfs_alloc_group.

We have two options:

1) Remove configfs_alloc_group/configfs_free_group as people will directly
use our new configfs_register_default_group that does the allocation.

2) Keep configfs_alloc/group/configfs_gree_group and use it here.

I'm in favour of 1) variant, as it makes little sense to have a wrapper over
kmalloc + config_group_init_type_name.

thanks,
Daniel.