Re: [PATCH v1 1/2] dt-bindings: usb: add TUSB73x0 PCIe

[Conor Dooley]

On Fri, Oct 04, 2024 at 05:31:04PM +0200, Francesco Dolcini wrote:
> Hello Conor,
> 
> On Fri, Oct 04, 2024 at 04:23:18PM +0100, Conor Dooley wrote:
> > On Fri, Oct 04, 2024 at 02:45:20PM +0200, Francesco Dolcini wrote:
> > > From: Parth Pancholi <parth.pancholi@xxxxxxxxxxx>
> > > 
> > > Add device tree bindings for TI's TUSB73x0 PCIe-to-USB 3.0 xHCI
> > > host controller. The controller supports software configuration
> > > through PCIe registers, such as controlling the PWRONx polarity
> > > via the USB control register (E0h).
> > > 
> > > Similar generic PCIe-based bindings can be found as qcom,ath11k-pci.yaml
> > > as an example.
> > > 
> > > Datasheet: https://www.ti.com/lit/ds/symlink/tusb7320.pdf
> > > Signed-off-by: Parth Pancholi <parth.pancholi@xxxxxxxxxxx>
> > > Signed-off-by: Francesco Dolcini <francesco.dolcini@xxxxxxxxxxx>
> > > ---
> > >  .../bindings/usb/ti,tusb73x0-pci.yaml         | 60 +++++++++++++++++++
> > >  1 file changed, 60 insertions(+)
> > >  create mode 100644 Documentation/devicetree/bindings/usb/ti,tusb73x0-pci.yaml
> > > 
> > > diff --git a/Documentation/devicetree/bindings/usb/ti,tusb73x0-pci.yaml b/Documentation/devicetree/bindings/usb/ti,tusb73x0-pci.yaml
> > > new file mode 100644
> > > index 000000000000..bcb619b08ad3
> > > --- /dev/null
> > > +++ b/Documentation/devicetree/bindings/usb/ti,tusb73x0-pci.yaml
> > > @@ -0,0 +1,60 @@
> > > +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
> > > +%YAML 1.2
> > > +---
> > > +$id: http://devicetree.org/schemas/usb/ti,tusb73x0-pci.yaml#
> > > +$schema: http://devicetree.org/meta-schemas/core.yaml#
> > > +
> > > +title: TUSB73x0 USB 3.0 xHCI Host Controller (PCIe)
> > > +
> > > +maintainers:
> > > +  - Francesco Dolcini <francesco.dolcini@xxxxxxxxxxx>
> > > +
> > > +description:
> > > +  TUSB73x0 USB 3.0 xHCI Host Controller via PCIe x1 Gen2 interface.
> > > +  The TUSB7320 supports up to two downstream ports, the TUSB7340 supports up
> > > +  to four downstream ports.
> > > +
> > > +properties:
> > > +  compatible:
> > > +    const: pci104C,8241
> > > +
> > > +  reg:
> > > +    maxItems: 1
> > > +
> > > +  ti,tusb7320-pwron-polarity-invert:
> > 
> > To me, "polarity-invert" makes less sense than calling this "active-high"
> > making the property a flag. active-low would then be the case where the
> > property is not provided. Given you don't make the property required,
> > what you've got here is effectively a flag anyway.
> 
> We had the same doubt when deciding which property name to propose, looking
> at the existing bindings it seemed that "polarity-invert" was more common.
> 
> FTR the datasheet explicetly name the signals with a # suffix (PWRON1#,
> PWRON2#, ...), they are defined as active-low by default.
> 
> With that said, if we prefer to have `ti,tusb7320-pwron-active-high`, I am 100%
> good with it.

I think "active-high" is more explicit about what it does, but I'm not
too bothered about it. Given it isn't a required property and the
absence of the property means active-low (for backwards compatibility
and alignment with the default hardware behaviour) the property is a
flag, not a boolean, so it does need a type change at the very least.

Cheers,
Conor.

Attachment: signature.asc
Description: PGP signature