linux-gpio.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Linus Walleij <linus.walleij@linaro.org>
To: Peter Ujfalusi <peter.ujfalusi@ti.com>
Cc: Bartosz Golaszewski <bgolaszewski@baylibre.com>,
	Rob Herring <robh+dt@kernel.org>,
	"open list:GPIO SUBSYSTEM" <linux-gpio@vger.kernel.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	Marek Szyprowski <m.szyprowski@samsung.com>,
	Mark Brown <broonie@kernel.org>,
	Maxime Ripard <mripard@kernel.org>,
	Philipp Zabel <p.zabel@pengutronix.de>,
	"open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS" 
	<devicetree@vger.kernel.org>
Subject: Re: [RFC 2/2] gpiolib: Support for (output only) shared GPIO line
Date: Fri, 22 Nov 2019 13:22:23 +0100	[thread overview]
Message-ID: <CACRpkdYt5P=GNc3EgHb-ry9fxMbXfpZd4FC=tuLqonNJKUM2wg@mail.gmail.com> (raw)
In-Reply-To: <20191120133409.9217-3-peter.ujfalusi@ti.com>

On Wed, Nov 20, 2019 at 2:34 PM Peter Ujfalusi <peter.ujfalusi@ti.com> wrote:

> This patch adds basic support for handling shared GPIO lines in the core.
> The line must be configured with a child node in DT.
> Based on the configuration the core will use different strategy to manage
> the shared line:
> refcounted low: Keep the line low as long as there is at least one low
>                 request is registered
> refcounted high: Keep the line high as long as there is at least one high
>                 request is registered
> pass through: all requests are allowed to go through without refcounting.
>
> The pass through mode is equivalent to how currently the
> GPIOD_FLAGS_BIT_NONEXCLUSIVE is handled.
>
> Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>

This is a good start! Some ideas on how I'd like this to develop.

>  drivers/gpio/gpiolib-of.c |  28 ++++++--
>  drivers/gpio/gpiolib.c    | 132 +++++++++++++++++++++++++++++++++++---

Please put this API under its own Kconfig option
and in its own file in
drivers/gpio/gpiolib-refcounted.c
local header in
drivers/gpio/gpiolib-refcounted.h
only built in if the appropriate Kconfig is selected.

Consumer header in
include/linux/gpio/reference-counted.h
And add external driver API to this last file.

> --- a/drivers/gpio/gpiolib-of.c

No commenting on this because as pointed out in the binding
patch I want this done by simply detecting the same GPIO
being referenced by several <&gpio N> phandles.

> diff --git a/drivers/gpio/gpiolib.h b/drivers/gpio/gpiolib.h
> index ca9bc1e4803c..0eec0857e3a8 100644
> --- a/drivers/gpio/gpiolib.h
> +++ b/drivers/gpio/gpiolib.h
> @@ -111,11 +111,18 @@ struct gpio_desc {
>  #define FLAG_PULL_UP    13     /* GPIO has pull up enabled */
>  #define FLAG_PULL_DOWN  14     /* GPIO has pull down enabled */
>  #define FLAG_BIAS_DISABLE    15        /* GPIO has pull disabled */
> +#define FLAG_IS_SHARED 16      /* GPIO is shared */

This is a good way of flagging that this is a refcounted GPIO
I would call it FLAG_IS_REFERENCE_COUNTED as it is
more precise to what it means.

> +#define FLAG_REFCOUNT_LOW 17   /* Shared GPIO is refcounted for raw low */
> +#define FLAG_REFCOUNT_HIGH 18  /* Shared GPIO is refcounted for raw high */

Do not put this here, keep it in the local refcounted GPIO
struct gpio_refcount_desc.

>         /* Connection label */
>         const char              *label;
>         /* Name of the GPIO */
>         const char              *name;
> +       /* Number of users of a shared GPIO */
> +       int                     shared_users;
> +       /* Reference counter for shared GPIO (low or high level) */
> +       int                     level_refcount;

We can't expand this struct for everyone on the planet like this.

In the local header

drivers/gpio/gpiolib-refcount.h create something like:

struct gpio_refcount_desc {
    struct gpio_desc *gd;
    int shared_users;
    int level_refcount;
};

This should be the opaque cookie returned to consumers of this new
refcounted API.

It defines the reference counted API as separate and optional from
the non-reference counted API, also using its own API.

The only effect on the core
gpiolib will be the single flag in struct gpio_desc; and some
calls into the shared code with stubs if the support for systems
that do not enable the reference counting API.

Yours,
Linus Walleij

  reply	other threads:[~2019-11-22 12:22 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-11-20 13:34 [RFC 0/2] gpiolib: Initial, basic support for shared GPIO lines Peter Ujfalusi
2019-11-20 13:34 ` [RFC 1/2] dt-bindings: gpio: Document shared GPIO line usage Peter Ujfalusi
2019-11-22 12:10   ` Linus Walleij
2019-11-22 13:36     ` Peter Ujfalusi
2019-11-28 10:06       ` Linus Walleij
2019-12-02 21:31         ` Peter Ujfalusi
2019-12-11  0:06           ` Linus Walleij
2019-12-03 23:51         ` Rob Herring
2019-12-10 23:54           ` Linus Walleij
2019-11-20 13:34 ` [RFC 2/2] gpiolib: Support for (output only) shared GPIO line Peter Ujfalusi
2019-11-22 12:22   ` Linus Walleij [this message]
2019-11-22 15:14     ` Peter Ujfalusi
2019-11-20 13:49 ` [RFC 0/2] gpiolib: Initial, basic support for shared GPIO lines Peter Ujfalusi

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to='CACRpkdYt5P=GNc3EgHb-ry9fxMbXfpZd4FC=tuLqonNJKUM2wg@mail.gmail.com' \
    --to=linus.walleij@linaro.org \
    --cc=bgolaszewski@baylibre.com \
    --cc=broonie@kernel.org \
    --cc=devicetree@vger.kernel.org \
    --cc=linux-gpio@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=m.szyprowski@samsung.com \
    --cc=mripard@kernel.org \
    --cc=p.zabel@pengutronix.de \
    --cc=peter.ujfalusi@ti.com \
    --cc=robh+dt@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).