linux-kernel - Re: [RFC 0/2] gpiolib: Initial, basic support for shared GPIO lines

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Wed, 20 Nov 2019 15:49:51 +0200
From:   Peter Ujfalusi <peter.ujfalusi@...com>
To:     <linus.walleij@...aro.org>, <bgolaszewski@...libre.com>,
        <robh+dt@...nel.org>
CC:     <linux-gpio@...r.kernel.org>, <linux-kernel@...r.kernel.org>,
        <m.szyprowski@...sung.com>, <broonie@...nel.org>,
        <mripard@...nel.org>, <p.zabel@...gutronix.de>,
        <devicetree@...r.kernel.org>
Subject: Re: [RFC 0/2] gpiolib: Initial, basic support for shared GPIO lines

Hi,

On 20/11/2019 15.34, Peter Ujfalusi wrote:
> Hi,
> 
> The initial support can replace all use of GPIOD_FLAGS_BIT_NONEXCLUSIVE if the
> shared GPIO is configured to follow pass through 'strategy' for the shared GPIO
> pin.
> 
> I have only implemented DT support.
> 
> With the shared gpio support one can choose between three different strategy for
> managing the shared gpio:
> 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.
> 
> Few shortcomings as of now:
> - can not handle different GPIO_ACTIVE_ on the user side, both the root GPIO
>   (which is shared) and clients must have the same GPIO_ACTIVE_ mode.
>   We are using common gpio_desc.
>   Like with GPIOD_FLAGS_BIT_NONEXCLUSIVE
> - refcounting counts _all_ 1/0 requests coming from the users of the shared
>   GPIO. This could cause issues if clients are using the gpiod API in unbalanced
>   way.
>   We would need to have separate tracking for each of the clients and agregate
>   the level they are asking for at any moment. Basically a new gpio-chip on top
>   of the real gpio pin can solve this.

I forgot to add an example for DT.

I have this for two pcm3168a codec, their RST pin is connected to
tca6416's p0 pin.
The codec's RST line is low active, so the gpio-shared is configured to
refcounted high to make sure that a codec is not placed in reset when it
does not want it.

&main_i2c3 {
	#address-cells = <1>;
	#size-cells = <0>;

	audio_exp: gpio@21 {
		compatible = "ti,tca6416";
		reg = <0x21>;
		gpio-controller;
		#gpio-cells = <2>;

		p00 {
			gpio-shared;
			gpios = <0 GPIO_ACTIVE_LOW>;
			output-high;
			refcounted-high;
			line-name = "CODEC RESET";
		};
	};

	pcm3168a_a: audio-codec@47 {
		compatible = "ti,pcm3168a";
		reg = <0x47>;
		#sound-dai-cells = <1>;

		reset-gpios = <&audio_exp 0 GPIO_ACTIVE_LOW>;
	};

	pcm3168a_b: audio-codec@46 {
		compatible = "ti,pcm3168a";
		reg = <0x46>;
		#sound-dai-cells = <1>;

		reset-gpios = <&audio_exp 0 GPIO_ACTIVE_LOW>;
	};
};


> 
> Regards,
> Peter
> ---
> Peter Ujfalusi (2):
>   dt-bindings: gpio: Document shared GPIO line usage
>   gpiolib: Support for (output only) shared GPIO line
> 
>  .../devicetree/bindings/gpio/gpio.txt         |  66 +++++++++
>  drivers/gpio/gpiolib-of.c                     |  28 +++-
>  drivers/gpio/gpiolib.c                        | 132 ++++++++++++++++--
>  drivers/gpio/gpiolib.h                        |  10 ++
>  4 files changed, 223 insertions(+), 13 deletions(-)
> 

- Péter

Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki.
Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki