Re: [PATCH 2/3] dt-bindings: remoteproc: Add AVM WASP

[Rob Herring <robh@kernel.org>]

On Mon, Feb 21, 2022 at 05:47:34PM +0100, Krzysztof Kozlowski wrote:
> On 21/02/2022 14:53, Daniel Kestrel wrote:
> > AVM Fritzbox router boards may contain an additional ATH79
> > based SoC that has the wifi cards connected.
> > This patch adds bindings for this remote processor.
> > 
> > Signed-off-by: Daniel Kestrel <kestrelseventyfour@gmail.com>
> > ---
> >  .../bindings/remoteproc/avm,wasp-rproc.yaml   | 93 +++++++++++++++++++
> >  1 file changed, 93 insertions(+)
> >  create mode 100644 Documentation/devicetree/bindings/remoteproc/avm,wasp-rproc.yaml
> > 
> > diff --git a/Documentation/devicetree/bindings/remoteproc/avm,wasp-rproc.yaml b/Documentation/devicetree/bindings/remoteproc/avm,wasp-rproc.yaml
> > new file mode 100644
> > index 000000000000..21f3bbcc4202
> > --- /dev/null
> > +++ b/Documentation/devicetree/bindings/remoteproc/avm,wasp-rproc.yaml
> > @@ -0,0 +1,93 @@
> > +# SPDX-License-Identifier: (GPL-2.0 OR BSD-2-Clause)
> > +%YAML 1.2
> > +---
> > +$id: http://devicetree.org/schemas/remoteproc/avm,wasp-rproc.yaml#
> > +$schema: http://devicetree.org/meta-schemas/core.yaml#
> > +
> > +title: AVM WASP processor controller bindings
> > +
> > +maintainers:
> > +  - Daniel Kestrel <kestrelseventyfour@gmail.com>
> > +
> > +description: |
> > +  This document defines the bindings for the remoteproc component that loads and
> > +  boots firmwares on the AVM Wireless Assistent Support Processor (WASP) SoC
> > +  that is attached to some AVM Fritzbox devices (3390, 3490, 5490, 5491, 7490).
> > +
> > +properties:
> > +  compatible:
> > +    const: avm,wasp
> > +
> > +  ath9k-firmware:
> > +    $ref: /schemas/types.yaml#/definitions/string
> > +    description: |
> > +      Should contain the name of the ath9k eeprom that is to be loaded from
> > +      the lantiq host flash. Wifi on the WASP SoC does not work without it.
> > +      The file should be located on the firmware search path.
> 
> Are you sure this is a property of hardware? It looks like runtime
> configuration parameter.

The standardish name for this is 'firmware-name'.

'name of the ath9k eeprom' is an odd description given there is no 
eeprom in this case. Where it is loaded from exactly is outside the 
scope of this binding.

> 
> > +
> > +  ath10k-caldata:
> > +    $ref: /schemas/types.yaml#/definitions/string
> > +    description: |
> > +      Should contain the name of the ath10k caldata that is to be loaded from
> > +      the lantiq host flash. Wifi on the WASP SoC does not work without it.
> > +      The file should be located on the firmware search path.
> 
> Same.

Ideally, 'firmware-name' would cover both cases and just provide a base 
name that the driver transforms into file names.

> 
> > +
> > +  wasp-netboot-firmware:
> > +    $ref: /schemas/types.yaml#/definitions/string
> > +    description: |
> > +      Should contain the name of the netboot firmware that is to be loaded
> > +      and started on the WASP SoC using mdio in order to be able to load
> > +      the initramfs image as a second stage.

initramfs is a Linux detail and should not be in binding.

> > +      The file should be located on the firmware search path.
> 
> Same.
> 
> > +
> > +  wasp-netboot-mdio:
> > +    $ref: /schemas/types.yaml#/definitions/phandle
> > +    description: Reference to the Lantiq GSWIP switch mdio.
> 
> Vendor prefix.
> 
> > +
> > +  wasp-initramfs-port:
> > +    $ref: /schemas/types.yaml#/definitions/phandle
> > +    description: Reference to the network port, where the WASP SoC is connected to.
> 
> Vendor prefix.
> 
> > +
> > +  wasp-initramfs-image:
> > +    $ref: /schemas/types.yaml#/definitions/string
> > +    description: |
> > +      Should contain the name of the initramfs linux image that is to be loaded
> > +      and started on the WASP SoC.
> > +      The file should be located on the firmware search path.
> 
> initramfs path looks even less like a property of hardware... If you
> change initramfs from CPIO to initrd or GZ, hardware changes as well?

And simply not how standard initramfs loading works. Boot menu files are 
how one gives the bootloader a location of initramfs file and chosen is 
how the kernel gets the memory location it was loaded to.

Rob