Re: [PATCH 2/6] dt-bindings: mtd: Rewrite gpio-control-nand in schema

[Miquel Raynal <miquel.raynal@bootlin.com>]

Hello Linus,

linus.walleij@linaro.org wrote on Wed, 08 Nov 2023 15:33:50 +0100:

> This creates a schema for GPIO controlled NAND. The txt
> schema was old and wrong.
> 
> Mark the old way of passing GPIOs in a long array as
> deprecated and encourage per-pin GPIO assignments with
> the named nnn-gpios phandles.
> 
> I was unable to re-use raw-nand-chip.yaml or even
> nand-chip.yaml: the reason is that they both assume
> that we have potentially several NAND chips with chip
> selects and thus enforce a node name "nand@0" etc,
> which doesn't quite work for this device.

But what about nand-controller.yaml? This driver is just about
emulating what a NAND controller would do with GPIOs, any NAND chip can
be wired, no?

> Since the GPIO controlled NAND is both a NAND controller
> and a NAND chip jitted together, 

Not really, it's just the controller part? I know for years
NAND controllers, ECC engines and NAND chips have been considered a
single hardware entity, but I believe this one is just about emulating
the host controller part.

> I have to modify the
> mtd.yaml nodename requirement to include nand-controller@
> as this is the nodename that this device should use.
> 
> Deprecate the custom "band-width" property in favor of
> "nand-bus-width".
> 
> Reported-by: Howard Harte <hharte@magicandroidapps.com>
> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
> ---
> Check the required section especially. Since there is no
> hardware default for anything when using GPIOs for this,
> I think all these parameters are compulsory.
> ---
>  .../devicetree/bindings/mtd/gpio-control-nand.txt  |  47 ------
>  .../devicetree/bindings/mtd/gpio-control-nand.yaml | 168 +++++++++++++++++++++
>  Documentation/devicetree/bindings/mtd/mtd.yaml     |   2 +-
>  3 files changed, 169 insertions(+), 48 deletions(-)
> 
> diff --git a/Documentation/devicetree/bindings/mtd/gpio-control-nand.txt b/Documentation/devicetree/bindings/mtd/gpio-control-nand.txt
> deleted file mode 100644
> index 486a17d533d7..000000000000
> --- a/Documentation/devicetree/bindings/mtd/gpio-control-nand.txt
> +++ /dev/null
> @@ -1,47 +0,0 @@
> -GPIO assisted NAND flash
> -
> -The GPIO assisted NAND flash uses a memory mapped interface to
> -read/write the NAND commands and data and GPIO pins for the control
> -signals.
> -
> -Required properties:
> -- compatible : "gpio-control-nand"
> -- reg : should specify localbus chip select and size used for the chip.  The
> -  resource describes the data bus connected to the NAND flash and all accesses
> -  are made in native endianness.
> -- #address-cells, #size-cells : Must be present if the device has sub-nodes
> -  representing partitions.
> -- gpios : Specifies the GPIO pins to control the NAND device.  The order of
> -  GPIO references is:  RDY, nCE, ALE, CLE, and nWP. nCE and nWP are optional.
> -
> -Optional properties:
> -- bank-width : Width (in bytes) of the device.  If not present, the width
> -  defaults to 1 byte.
> -- chip-delay : chip dependent delay for transferring data from array to
> -  read registers (tR).  If not present then a default of 20us is used.
> -- gpio-control-nand,io-sync-reg : A 64-bit physical address for a read
> -  location used to guard against bus reordering with regards to accesses to
> -  the GPIO's and the NAND flash data bus.  If present, then after changing
> -  GPIO state and before and after command byte writes, this register will be
> -  read to ensure that the GPIO accesses have completed.
> -
> -The device tree may optionally contain sub-nodes describing partitions of the
> -address space. See partition.txt for more detail.
> -
> -Examples:
> -
> -gpio-nand@1,0 {
> -	compatible = "gpio-control-nand";
> -	reg = <1 0x0000 0x2>;
> -	#address-cells = <1>;
> -	#size-cells = <1>;
> -	gpios = <&banka 1 0>,	/* RDY */
> -		<0>, 		/* nCE */
> -		<&banka 3 0>, 	/* ALE */
> -		<&banka 4 0>, 	/* CLE */
> -		<0>;		/* nWP */
> -
> -	partition@0 {
> -	...
> -	};
> -};
> diff --git a/Documentation/devicetree/bindings/mtd/gpio-control-nand.yaml b/Documentation/devicetree/bindings/mtd/gpio-control-nand.yaml
> new file mode 100644
> index 000000000000..5b30ee7ad4a5
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/mtd/gpio-control-nand.yaml
> @@ -0,0 +1,168 @@
> +# SPDX-License-Identifier: GPL-2.0
> +%YAML 1.2
> +---
> +$id: http://devicetree.org/schemas/mtd/gpio-control-nand.yaml#
> +$schema: http://devicetree.org/meta-schemas/core.yaml#
> +
> +title: NAND memory exclusively connected to GPIO lines
> +
> +maintainers:
> +  - Linus Walleij <linus.walleij@linaro.org>
> +
> +description: |
> +  It is possible to connect a NAND flash memory without any
> +  dedicated NAND controller hardware, using just general purpose
> +  I/O (GPIO) pins. This will not be fast, but it will work.
> +  The address and data lines of the chip will still need to be
> +  connected so that the contents of a NAND page can be
> +  memory-mapped and accessed after the special lines are toggled
> +  by GPIO.
> +
> +# The GPIO controlled NAND has wires going directly to one single
> +# NAND chip, so it is both a nand controller and a nand chip at
> +# the same time, but it does not have things such as chip select
> +# since the use is hammered down to one single chip only.
> +# There is no point for the chip itself to be a subnode of the
> +# controller so the raw NAND chip properties are added right into
> +# the controller node like this.

I kind of disagree here, this "piece of software" only replaces a NAND
controller. You always need a NAND chip in front of it, and that's a
specific piece of hardware anyway. Or maybe I don't understand the
hardware behind? (truly not impossible)

> +
> +allOf:
> +  - $ref: mtd.yaml#
> +
> +properties:
> +  $nodename:
> +    pattern: "^(nand|nand-controller)@[a-f0-9]+$"
> +
> +  compatible:
> +    const: gpio-control-nand
> +
> +  reg:
> +    description: |
> +      This should specify the address where the NAND page currently
> +      accessed gets memory-mapped, and the size of the page. Usually
> +      this will be the same as the page size of the NAND.

This is definitely a host controller parameter. Even if the hardware
only supports a single NAND chip, I believe it should be described as a
subnode with a dummy "reg = <0>;" property.

Thanks,
Miquèl