Re: [PATCH v2 06/10] media: dt-bindings: Document SAMA5D4 VDEC bindings

[Emil Velikov <emil.l.velikov@gmail.com>]

On Fri, 26 Mar 2021 at 14:29, Ezequiel Garcia <ezequiel@collabora.com> wrote:
>
> Hi Emil,
>
> On Thu, 2021-03-25 at 14:14 +0000, Emil Velikov wrote:
> > On Wed, 24 Mar 2021 at 15:17, Rob Herring <robh@kernel.org> wrote:
> > >
> > > On Thu, Mar 11, 2021 at 03:40:51PM +0000, Emil Velikov wrote:
> > > > From: Emil Velikov <emil.velikov@collabora.com>
> > > >
> > > > Add devicetree binding documentation for the Hantro G1/G2 VDEC on
> > > > the Microchip SAMAS5D4 SoC.
> > > >
> > > > Cc: Rob Herring <robh+dt@kernel.org>
> > > > Cc: Frank Rowand <frowand.list@gmail.com>
> > > > Cc: devicetree@vger.kernel.org>
> > > > Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
> > > > ---
> > > > v2
> > > >  - Newly introduced
> > > >  - s/Atmel/Microchip/ (Nicolas)
> > > >  - Drop leading 0 in node name/address
> > > > ---
> > > >  .../media/microchip,sama5d4-vdec.yaml         | 59 +++++++++++++++++++
> > > >  1 file changed, 59 insertions(+)
> > > >  create mode 100644 Documentation/devicetree/bindings/media/microchip,sama5d4-vdec.yaml
> > > >
> > > > diff --git a/Documentation/devicetree/bindings/media/microchip,sama5d4-vdec.yaml b/Documentation/devicetree/bindings/media/microchip,sama5d4-
> > > > vdec.yaml
> > > > new file mode 100644
> > > > index 000000000000..9cb2c0295d54
> > > > --- /dev/null
> > > > +++ b/Documentation/devicetree/bindings/media/microchip,sama5d4-vdec.yaml
> > > > @@ -0,0 +1,59 @@
> > > > +# SPDX-License-Identifier: (GPL-2.0 OR BSD-2-Clause)
> > > > +
> > > > +%YAML 1.2
> > > > +---
> > > > +$id: "http://devicetree.org/schemas/media/microchip,sama5d4-vdec.yaml#"
> > > > +$schema: "http://devicetree.org/meta-schemas/core.yaml#"
> > > > +
> > > > +title: Hantro G1 VPU codec implemented on Microchip SAMA5D4 SoCs
> > > > +
> > > > +maintainers:
> > > > +  - Emil Velikov <emil.velikov@collabora.com>
> > > > +
> > > > +description:
> > > > +  Hantro G1 video decode accelerator present on Microchip SAMA5D4 SoCs.
> > > > +
> > > > +properties:
> > > > +  compatible:
> > > > +    const: microchip,sama5d4-vdec
> > > > +
> > > > +  reg:
> > > > +    maxItems: 1
> > > > +
> > > > +  interrupts:
> > > > +    maxItems: 1
> > > > +
> > > > +  interrupt-names:
> > > > +    items:
> > > > +      - const: vdec
> > >
> > > Why do you need a name? *-names are used to distinguish multiple entries
> > > and don't add anything if only a single entry.
> > >
> > On one hand the names are used to describe the hardware - the SAMA
> > board uses designated clock and interrupt lines - both called "vdec".
> > Additionally the names are ultimately required by the underlying API -
> > platform_get_irq_byname and  devm_clk_bulk_get respectively.
> > How can we get the respective entries without the name?
> >
>
> There are APIs to get the resource by index as opposed to by name.
> E.g. platform_get_irq.
>
> However, you are confusing here internal kernel implementation
> with devicetree bindings. The latter is just a representation
> of the hardware.
>
My train of thought was that if we're missing API to fetch the entries
via non-name means, then there's a reasonable precedent to expose the
name. Clearly I missed the API, so my argument is just off.

> > Skimming through the existing dts file -
> > arch/arm/boot/dts/sama5d4.dtsi and other dts files - there are lots of
> > examples where the device tree binding contains the name for a single
> > clock/interrupt.
> >
>
>
> Note that dtsi and dts are the device-tree files, not the bindings.
> And even if you find examples, the fact that there exist examples
> doesn't mean it's the right thing to do.
>
Fully agreed. Thanks for the correction and hints.

The v3, just submitted, lacks the names all together.

-Emil