[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1507555507.16356.50.camel@baylibre.com>
Date: Mon, 09 Oct 2017 15:25:07 +0200
From: Jerome Brunet <jbrunet@...libre.com>
To: Neil Armstrong <narmstrong@...libre.com>,
Linus Walleij <linus.walleij@...aro.org>,
Kevin Hilman <khilman@...libre.com>,
Carlo Caione <carlo@...one.org>
Cc: linux-gpio@...r.kernel.org, linux-amlogic@...ts.infradead.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH 2/2] pinctrl: meson: rework pinmux ops
On Mon, 2017-10-09 at 13:42 +0200, Neil Armstrong wrote:
> > +/* Common pmx functions */
> > +int meson_pmx_get_funcs_count(struct pinctrl_dev *pcdev);
> > +const char *meson_pmx_get_func_name(struct pinctrl_dev *pcdev,
> > + unsigned selector);
> > +int meson_pmx_get_groups(struct pinctrl_dev *pcdev,
> > + unsigned selector,
> > + const char * const **groups,
> > + unsigned * const num_groups);
>
> Maybe the naming of the common functions should be changed to something
> generic
> like meson_get_functions_name and meson_get_function_groups and leave "pmx" to
> the
> first version pinmux control implementation.
> Same for the ops, meson_pinmux_ops would be better.
I was thinking that the naming convention around this might be confusing. Thx
for the confirmation !
However, I think "pmx" was intially meant to denote a pinmux function, since
pinconf, gpio and pinmux are all mixed in pinctrl-meson.c. I'd prefer to keep
this "pmx" for this reason.
Maybe the SoC specific bits should be named after the first SoC supporting it:
* files: pinctrl-meson-pmx.* => pinctrl-meson8-pmx.*
* functions: meson_pmx_* => meson8_pmx_*
What do you think ?
Powered by blists - more mailing lists