Re: [PATCH 2/2] debugfs: make return value of all debugfs helpers consistent

[Greg Kroah-Hartman]

On Mon, Apr 15, 2019 at 04:29:59PM -0700, Life is hard, and then you die wrote:
> 
>   Hi Greg,
> 
> On Mon, Apr 15, 2019 at 10:48:44AM +0200, Greg Kroah-Hartman wrote:
> > On Mon, Apr 15, 2019 at 01:25:06AM -0700, Ronald Tschalär wrote:
> > > Since commit ff9fb72bc077 ("debugfs: return error values, not NULL")
> > > almost all the debugfs helpers have stopped returning NULL. The lone
> > > holdeout was debugfs_create_u32_array(). So fix that.
> > > 
> > > Signed-off-by: Ronald Tschalär <ronald@xxxxxxxxxxxxx>
> > > ---
> > >  fs/debugfs/file.c | 6 +++---
> > >  1 file changed, 3 insertions(+), 3 deletions(-)
> > > 
> > > diff --git a/fs/debugfs/file.c b/fs/debugfs/file.c
> > > index ddd708b09fa1..bb706d073782 100644
> > > --- a/fs/debugfs/file.c
> > > +++ b/fs/debugfs/file.c
> > > @@ -999,8 +999,8 @@ static const struct file_operations u32_array_fops = {
> > >   * Once array is created its size can not be changed.
> > >   *
> > >   * The function returns a pointer to dentry on success. If an error occurs,
> > > - * %ERR_PTR(-ERROR) or NULL will be returned. If debugfs is not enabled in
> > > - * the kernel, the value %ERR_PTR(-ENODEV) will be returned.
> > > + * %ERR_PTR(-ERROR) will be returned. If debugfs is not enabled in the kernel,
> > > + * the value %ERR_PTR(-ENODEV) will be returned.
> > >   */
> > >  struct dentry *debugfs_create_u32_array(const char *name, umode_t mode,
> > >  					    struct dentry *parent,
> > > @@ -1009,7 +1009,7 @@ struct dentry *debugfs_create_u32_array(const char *name, umode_t mode,
> > >  	struct array_data *data = kmalloc(sizeof(*data), GFP_KERNEL);
> > >  
> > >  	if (data == NULL)
> > > -		return NULL;
> > > +		return ERR_PTR(-ENOMEM);
> > >  
> > >  	data->array = array;
> > >  	data->elements = elements;
> > 
> > There is only one caller of this function in the kernel now, and it does
> > not even care about the return value at all, so we should just remove
> > the return value entirely as that's the easiest and best thing to do
> > here.
> 
> Interesting argument: since this is a helper/library function, and
> therefore potentially used in the future by others, it seems to me
> that consistency with the other functions and providing error feedback
> would be important.

Not if you never actually use the return value for anything :)

> > I was going to start doing this slowly over time, but as you are
> > touching the function now, might as well do it here :)
> 
> Are you saying the plan is to make all these helpers return void?

Yes, no caller should do anything "different" based on a return value of
a debugfs call.  Note, sometimes you do want to save off dentries for
some files that you later remove, but that's it.

thanks,

greg k-h