[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20230310144840.5203fe50@xps-13>
Date: Fri, 10 Mar 2023 14:48:40 +0100
From: Miquel Raynal <miquel.raynal@...tlin.com>
To: Rafał Miłecki <zajec5@...il.com>
Cc: Rob Herring <robh+dt@...nel.org>,
Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
Srinivas Kandagatla <srinivas.kandagatla@...aro.org>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Michael Walle <michael@...le.cc>, devicetree@...r.kernel.org,
linux-mtd@...ts.infradead.org, linux-kernel@...r.kernel.org,
Rafał Miłecki <rafal@...ecki.pl>
Subject: Re: [PATCH] dt-bindings: nvmem: allow MTD to be explicitly an NVMEM
provider
Hi Rafał,
zajec5@...il.com wrote on Fri, 10 Mar 2023 11:53:30 +0100:
> From: Rafał Miłecki <rafal@...ecki.pl>
>
> There are a lot of devices with NVMEM content stored in MTD devices in
> relevant partitions. Add a DT binding for marking such partitions.
>
> Note: Linux already treats every MTD partition as NVMEM provider so in
> general it doesn't need to care about this binding. It's meant just to
> make DT clearer in describing hardware.
>
> Signed-off-by: Rafał Miłecki <rafal@...ecki.pl>
> ---
> As explained in commit body this isn't really needed for Linux. I
> thought it'd be a small nice addition for writing clear DTS files.
> ---
> .../devicetree/bindings/nvmem/mtd.yaml | 52 +++++++++++++++++++
> 1 file changed, 52 insertions(+)
> create mode 100644 Documentation/devicetree/bindings/nvmem/mtd.yaml
>
> diff --git a/Documentation/devicetree/bindings/nvmem/mtd.yaml b/Documentation/devicetree/bindings/nvmem/mtd.yaml
> new file mode 100644
> index 000000000000..7435b2803cf9
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/nvmem/mtd.yaml
> @@ -0,0 +1,52 @@
> +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
> +%YAML 1.2
> +---
> +$id: http://devicetree.org/schemas/nvmem/mtd.yaml#
> +$schema: http://devicetree.org/meta-schemas/core.yaml#
> +
> +title: MTD access based NVMEM
> +
> +description: |
> + MTD partitions can be NVMEM providers. This binding allows explicitly marking
> + such partitions.
We already have that, it's nvmem-cell? I understand what you want to
do, but I think it suffers from a common problem, see below.
> + The exact way of handling MTD partition content (NVMEM cells) should be
> + described using a proper NVMEM layout.
Ok so I believe this is another solution for the layout offset proposed
by Michael. Except that it only fixes it for mtd. I think I would
prefer the former solution which handles all nvmem cases.
> +
> +maintainers:
> + - Rafał Miłecki <rafal@...ecki.pl>
> +
> +allOf:
> + - $ref: nvmem.yaml#
> + - $ref: /schemas/mtd/partitions/partition.yaml#
> +
> +properties:
> + compatible:
> + const: mtd-nvmem
> +
> + reg:
> + maxItems: 1
> +
> +required:
> + - reg
> +
> +unevaluatedProperties: false
> +
> +examples:
> + - |
> + partitions {
> + compatible = "fixed-partitions";
> + #address-cells = <1>;
> + #size-cells = <1>;
> +
> + partition@0 {
> + compatible = "mtd-nvmem";
Actually there has been valid push-back from devlink gurus against stale
compatibles (on a device-driver point of view) like that. Maybe
something like this instead:
partition@x{
<mtd-nvmem-property>
?
> + reg = <0x0 0x40000>;
> + label = "device-data";
> +
> + nvmem-layout {
> + /* Just a dummy example: Kontron can be found on OTP actually */
> + compatible = "kontron,sl28-vpd";
The Onie tlv compatible would perfectly apply here.
> + };
> + };
> + };
Thanks,
Miquèl
Powered by blists - more mailing lists