2014-10-27 15:27:40

by Torsten Fleischer

[permalink] [raw]
Subject: [PATCH 1/1] spi: spi-gpio: Add dt support for a single device with no chip select

From: Torsten Fleischer <[email protected]>

Use an empty 'cs-gpios' property if a single device is connected which
has no chip select.
This refers to the discussion at https://lkml.org/lkml/2013/11/13/62.

Signed-off-by: Torsten Fleischer <[email protected]>
---
Documentation/devicetree/bindings/spi/spi-gpio.txt | 6 ++++--
drivers/spi/spi-gpio.c | 10 +++++++---
2 files changed, 11 insertions(+), 5 deletions(-)

diff --git a/Documentation/devicetree/bindings/spi/spi-gpio.txt b/Documentation/devicetree/bindings/spi/spi-gpio.txt
index 8a824be..ad8795f 100644
--- a/Documentation/devicetree/bindings/spi/spi-gpio.txt
+++ b/Documentation/devicetree/bindings/spi/spi-gpio.txt
@@ -8,8 +8,10 @@ Required properties:
- gpio-sck: GPIO spec for the SCK line to use
- gpio-miso: GPIO spec for the MISO line to use
- gpio-mosi: GPIO spec for the MOSI line to use
- - cs-gpios: GPIOs to use for chipselect lines
- - num-chipselects: number of chipselect lines
+ - cs-gpios: GPIOs to use for chipselect lines. Leave this property empty if
+ a single device with no chip select is connected.
+ - num-chipselects: Number of chipselect lines. Has to be "1" even if a single
+ device with no chip select is connected.

Example:

diff --git a/drivers/spi/spi-gpio.c b/drivers/spi/spi-gpio.c
index 9f59553..b343818a 100644
--- a/drivers/spi/spi-gpio.c
+++ b/drivers/spi/spi-gpio.c
@@ -461,9 +461,13 @@ static int spi_gpio_probe(struct platform_device *pdev)
* property of the node.
*/

- for (i = 0; i < SPI_N_CHIPSEL; i++)
- spi_gpio->cs_gpios[i] =
- of_get_named_gpio(np, "cs-gpios", i);
+ if ((SPI_N_CHIPSEL == 1) &&
+ (of_gpio_named_count(np, "cs-gpios") == 0))
+ spi_gpio->cs_gpios[0] = SPI_GPIO_NO_CHIPSELECT;
+ else
+ for (i = 0; i < SPI_N_CHIPSEL; i++)
+ spi_gpio->cs_gpios[i] =
+ of_get_named_gpio(np, "cs-gpios", i);
}
#endif

--
1.8.4.5


2014-10-27 17:31:15

by Mark Rutland

[permalink] [raw]
Subject: Re: [PATCH 1/1] spi: spi-gpio: Add dt support for a single device with no chip select

On Mon, Oct 27, 2014 at 03:27:17PM +0000, Torsten Fleischer wrote:
> From: Torsten Fleischer <[email protected]>
>
> Use an empty 'cs-gpios' property if a single device is connected which
> has no chip select.
> This refers to the discussion at https://lkml.org/lkml/2013/11/13/62.
>
> Signed-off-by: Torsten Fleischer <[email protected]>
> ---
> Documentation/devicetree/bindings/spi/spi-gpio.txt | 6 ++++--
> drivers/spi/spi-gpio.c | 10 +++++++---
> 2 files changed, 11 insertions(+), 5 deletions(-)
>
> diff --git a/Documentation/devicetree/bindings/spi/spi-gpio.txt b/Documentation/devicetree/bindings/spi/spi-gpio.txt
> index 8a824be..ad8795f 100644
> --- a/Documentation/devicetree/bindings/spi/spi-gpio.txt
> +++ b/Documentation/devicetree/bindings/spi/spi-gpio.txt
> @@ -8,8 +8,10 @@ Required properties:
> - gpio-sck: GPIO spec for the SCK line to use
> - gpio-miso: GPIO spec for the MISO line to use
> - gpio-mosi: GPIO spec for the MOSI line to use
> - - cs-gpios: GPIOs to use for chipselect lines
> - - num-chipselects: number of chipselect lines
> + - cs-gpios: GPIOs to use for chipselect lines. Leave this property empty if
> + a single device with no chip select is connected.
> + - num-chipselects: Number of chipselect lines. Has to be "1" even if a single
> + device with no chip select is connected.

Why not just have num-chipselectes = <0>, and no cs-gpios property?

That would match the HW, and is unambigious as the author has to state
explicitly that there are no chipselects.

Mark.

>
> Example:
>
> diff --git a/drivers/spi/spi-gpio.c b/drivers/spi/spi-gpio.c
> index 9f59553..b343818a 100644
> --- a/drivers/spi/spi-gpio.c
> +++ b/drivers/spi/spi-gpio.c
> @@ -461,9 +461,13 @@ static int spi_gpio_probe(struct platform_device *pdev)
> * property of the node.
> */
>
> - for (i = 0; i < SPI_N_CHIPSEL; i++)
> - spi_gpio->cs_gpios[i] =
> - of_get_named_gpio(np, "cs-gpios", i);
> + if ((SPI_N_CHIPSEL == 1) &&
> + (of_gpio_named_count(np, "cs-gpios") == 0))
> + spi_gpio->cs_gpios[0] = SPI_GPIO_NO_CHIPSELECT;
> + else
> + for (i = 0; i < SPI_N_CHIPSEL; i++)
> + spi_gpio->cs_gpios[i] =
> + of_get_named_gpio(np, "cs-gpios", i);
> }
> #endif
>
> --
> 1.8.4.5
>
> --
> To unsubscribe from this list: send the line "unsubscribe devicetree" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>

2014-10-30 06:52:13

by Torsten Fleischer

[permalink] [raw]
Subject: Re: [PATCH 1/1] spi: spi-gpio: Add dt support for a single device with no chip select

2014-10-27 18:30 GMT+01:00 Mark Rutland <[email protected]>:
> On Mon, Oct 27, 2014 at 03:27:17PM +0000, Torsten Fleischer wrote:
>> From: Torsten Fleischer <[email protected]>
>>
>> Use an empty 'cs-gpios' property if a single device is connected which
>> has no chip select.
>> This refers to the discussion at https://lkml.org/lkml/2013/11/13/62.
>>
>> Signed-off-by: Torsten Fleischer <[email protected]>
>> ---
>> Documentation/devicetree/bindings/spi/spi-gpio.txt | 6 ++++--
>> drivers/spi/spi-gpio.c | 10 +++++++---
>> 2 files changed, 11 insertions(+), 5 deletions(-)
>>
>> diff --git a/Documentation/devicetree/bindings/spi/spi-gpio.txt b/Documentation/devicetree/bindings/spi/spi-gpio.txt
>> index 8a824be..ad8795f 100644
>> --- a/Documentation/devicetree/bindings/spi/spi-gpio.txt
>> +++ b/Documentation/devicetree/bindings/spi/spi-gpio.txt
>> @@ -8,8 +8,10 @@ Required properties:
>> - gpio-sck: GPIO spec for the SCK line to use
>> - gpio-miso: GPIO spec for the MISO line to use
>> - gpio-mosi: GPIO spec for the MOSI line to use
>> - - cs-gpios: GPIOs to use for chipselect lines
>> - - num-chipselects: number of chipselect lines
>> + - cs-gpios: GPIOs to use for chipselect lines. Leave this property empty if
>> + a single device with no chip select is connected.
>> + - num-chipselects: Number of chipselect lines. Has to be "1" even if a single
>> + device with no chip select is connected.
>
> Why not just have num-chipselectes = <0>, and no cs-gpios property?
>
> That would match the HW, and is unambigious as the author has to state
> explicitly that there are no chipselects.

Thanks for the objection. I agree to your proposal. This is a clearer
way to describe such
a hardware. I will send an updated patch.

Torsten