2018-08-15 20:27:13

by David Bauer

[permalink] [raw]
Subject: [PATCH] gpio: 74x164: add lines-initial-states property

This adds the ability to define the initial state of each output line on
device probe.

Signed-off-by: David Bauer <[email protected]>
---
Documentation/devicetree/bindings/gpio/gpio-74x164.txt | 5 +++++
drivers/gpio/gpio-74x164.c | 3 +++
2 files changed, 8 insertions(+)

diff --git a/Documentation/devicetree/bindings/gpio/gpio-74x164.txt b/Documentation/devicetree/bindings/gpio/gpio-74x164.txt
index 2a97553d8d76..580b18065ad3 100644
--- a/Documentation/devicetree/bindings/gpio/gpio-74x164.txt
+++ b/Documentation/devicetree/bindings/gpio/gpio-74x164.txt
@@ -14,6 +14,11 @@ Required properties:

Optional properties:
- enable-gpios: GPIO connected to the OE (Output Enable) pin.
+- lines-initial-states: Bitmask that specifies the initial state of
+ each line. When a bit is set to zero, the corresponding output line
+ is initialized LOW. When a bit is set to one, the corresponding
+ output line is initialized HIGH. In case this property is not
+ defined, all lines will be initialized as LOW.

Example:

diff --git a/drivers/gpio/gpio-74x164.c b/drivers/gpio/gpio-74x164.c
index fb7b620763a2..275310a0a538 100644
--- a/drivers/gpio/gpio-74x164.c
+++ b/drivers/gpio/gpio-74x164.c
@@ -150,6 +150,9 @@ static int gen_74x164_probe(struct spi_device *spi)
chip->gpio_chip.parent = &spi->dev;
chip->gpio_chip.owner = THIS_MODULE;

+ of_property_read_u8_array(spi->dev.of_node, "lines-initial-states",
+ chip->buffer, chip->registers);
+
mutex_init(&chip->lock);

ret = __gen_74x164_write_config(chip);
--
2.18.0



2018-08-16 15:28:22

by Linus Walleij

[permalink] [raw]
Subject: Re: [PATCH] gpio: 74x164: add lines-initial-states property

Hi David,

On Wed, Aug 15, 2018 at 10:19 PM David Bauer <[email protected]> wrote:

> This adds the ability to define the initial state of each output line on
> device probe.
>
> Signed-off-by: David Bauer <[email protected]>

(...)
> Optional properties:
> - enable-gpios: GPIO connected to the OE (Output Enable) pin.
> +- lines-initial-states: Bitmask that specifies the initial state of
> + each line. When a bit is set to zero, the corresponding output line
> + is initialized LOW. When a bit is set to one, the corresponding
> + output line is initialized HIGH. In case this property is not
> + defined, all lines will be initialized as LOW.

This sounds like something that should be generic, and not use
a bitmask, but offsets. It should work even if the number of
GPIOs from the chip is > 32.

Is the usecase different from hogs?
See Documentation/devicetree/bindings/gpio.txt

There has been extensive discussion about supporting initial values
with something similar to hogs, but I haven't got anything ACKed
by the DT maintainers so it has kind of stalled.

I would make sure to both make it generic, get ACK from the DT
mainatiners, and make sure to implement it in
drivers/gpio/gpiolib-of.c and not locally in drivers.

Yours,
Linus Walleij

2018-08-17 15:11:39

by Rob Herring (Arm)

[permalink] [raw]
Subject: Re: [PATCH] gpio: 74x164: add lines-initial-states property

Hi, this email is from Rob's (experimental) review bot. I found a couple
of common problems with your patch. Please see below.

On Wed, 15 Aug 2018 22:18:54 +0200, David Bauer wrote:
> This adds the ability to define the initial state of each output line on
> device probe.
>
> Signed-off-by: David Bauer <[email protected]>

The preferred subject prefix is "dt-bindings: <binding dir>: ...".

> ---
> Documentation/devicetree/bindings/gpio/gpio-74x164.txt | 5 +++++
> drivers/gpio/gpio-74x164.c | 3 +++
> 2 files changed, 8 insertions(+)
>

DT bindings (including binding headers) should be a separate patch. See
Documentation/devicetree/bindings/submitting-patches.txt.

2018-08-19 23:58:52

by David Bauer

[permalink] [raw]
Subject: Re: [PATCH] gpio: 74x164: add lines-initial-states property

Hi Linus,

On 8/16/18 10:11 AM, Linus Walleij wrote:
> This sounds like something that should be generic, and not use
> a bitmask, but offsets. It should work even if the number of
> GPIOs from the chip is > 32.
>
> Is the usecase different from hogs?
> See Documentation/devicetree/bindings/gpio.txt

Thanks for pointing that out. Indeed for my use-case (Asserting single
output to be high on driver probe) hogs are are sufficient solution.

Best wishes
David