Re: [PATCH 8/8] ALSA: control: arrange snd_ctl_new() as a local function

[Takashi Iwai <tiwai@suse.de>]

At Thu, 12 Feb 2015 21:45:24 +0900,
Takashi Sakamoto wrote:
> 
> On Feb 11 2015 22:04, Takashi Iwai wrote:
> > At Wed, 11 Feb 2015 13:49:29 +0100,
> > Lars-Peter Clausen wrote:
> >>
> >> On 02/11/2015 11:37 AM, Takashi Sakamoto wrote:
> >> [...]
> >>> -/**
> >>> - * snd_ctl_new - create a control instance from the template
> >>> - * @control: the control template
> >>> - * @access: the default control access
> >>> - *
> >>> - * Allocates a new struct snd_kcontrol instance and copies the given template
> >>> - * to the new instance. It does not copy volatile data (access).
> >>> - *
> >>> - * Return: The pointer of the new instance, or %NULL on failure.
> >>> - */
> >>> -static struct snd_kcontrol *snd_ctl_new(struct snd_kcontrol *control,
> >>> -					unsigned int access)
> >>> +static struct snd_kcontrol *ctl_new(struct snd_kcontrol *control,
> >>
> >> I'd prefer to keep both the documentation as well as the 'snd_' prefix. 
> >> Maybe just replace the '/**' with '/*' to prevent it from appearing in the 
> >> public API documentation.
> > 
> > Yes, agreed.  There is no strict rule about snd_ prefix for
> > non-exported objects.  We prefer not having snd_ prefix for local
> > stuff in general just for readability.  But in this case, it doesn't
> > improve so much.
> 
> Hm. I don't disagree with remaining these comments.
> 
> ...But I have another issue. In my final patch in another series, I
> change this function drastically. Then, should I spend a bit time to
> revise them even if this is just a local function?

Yes, please.


Takashi