linux-kernel - Re: [PATCH] dt-bindings: phy: qcom,sc8280xp-qmp-usb43dp: do not reference whole usb-switch.yaml

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20251001163139.GA1877961-robh@kernel.org>
Date: Wed, 1 Oct 2025 11:31:39 -0500
From: Rob Herring <robh@...nel.org>
To: Dmitry Baryshkov <dmitry.baryshkov@....qualcomm.com>
Cc: Neil Armstrong <neil.armstrong@...aro.org>,
	Vinod Koul <vkoul@...nel.org>,
	Kishon Vijay Abraham I <kishon@...nel.org>,
	Krzysztof Kozlowski <krzk+dt@...nel.org>,
	Conor Dooley <conor+dt@...nel.org>,
	Konrad Dybcio <konrad.dybcio@....qualcomm.com>,
	Krzysztof Kozlowski <krzk@...nel.org>,
	linux-arm-msm@...r.kernel.org, linux-phy@...ts.infradead.org,
	devicetree@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] dt-bindings: phy: qcom,sc8280xp-qmp-usb43dp: do not
 reference whole usb-switch.yaml

On Tue, Sep 30, 2025 at 10:10:25PM +0300, Dmitry Baryshkov wrote:
> On Fri, Sep 05, 2025 at 12:55:33PM -0500, Rob Herring wrote:
> > On Tue, Sep 02, 2025 at 06:10:05PM +0200, Neil Armstrong wrote:
> > > Both bindings describe a different layout of the ports properties,
> > > leading to errors when validating DT using this PHY bindings as
> > > reported by Rob Herring.
> > > 
> > > Reported-by: Rob Herring <robh@...nel.org>
> > > Closes: https://lore.kernel.org/all/175462129176.394940.16810637795278334342.robh@kernel.org/
> > > Fixes: 3bad7fe22796 ("dt-bindings: phy: qcom,sc8280xp-qmp-usb43dp: Reference usb-switch.yaml to allow mode-switch")
> > > Signed-off-by: Neil Armstrong <neil.armstrong@...aro.org>
> > > ---
> > >  .../devicetree/bindings/phy/qcom,sc8280xp-qmp-usb43dp-phy.yaml    | 8 +++++---
> > >  1 file changed, 5 insertions(+), 3 deletions(-)
> > > 
> > > diff --git a/Documentation/devicetree/bindings/phy/qcom,sc8280xp-qmp-usb43dp-phy.yaml b/Documentation/devicetree/bindings/phy/qcom,sc8280xp-qmp-usb43dp-phy.yaml
> > > index c8bc512df08b5694c8599f475de78679a4438449..5005514d7c3a1e4a8893883497fd204bc04e12be 100644
> > > --- a/Documentation/devicetree/bindings/phy/qcom,sc8280xp-qmp-usb43dp-phy.yaml
> > > +++ b/Documentation/devicetree/bindings/phy/qcom,sc8280xp-qmp-usb43dp-phy.yaml
> > > @@ -73,8 +73,11 @@ properties:
> > >      description:
> > >        See include/dt-bindings/phy/phy-qcom-qmp.h
> > >  
> > > -  mode-switch: true
> > > -  orientation-switch: true
> > > +  mode-switch:
> > > +    $ref: /schemas/usb/usb-switch.yaml#properties/mode-switch
> > > +
> > > +  orientation-switch:
> > > +    $ref: /schemas/usb/usb-switch.yaml#properties/orientation-switch

Looking at this again, this isn't even correct and I don't think it 
works. It's missing a '/' and  should be ...#/properties/... to be a 
valid json pointer.

I thought we checked this...

> > 
> > This is a pattern we try to avoid with references at a property level. I 
> > prefer you make port and ports not required in usb-switch.yaml.
> 
> But this solution is also not perfect. E.g.
> Documentation/devicetree/bindings/phy/fsl,imx8mq-usb-phy.yaml should
> only allow the orienttion-switch property, while using
> allOf:$ref:usb-switch.yaml allows all three (orientation-switch,
> mode-switch, retimer-switch).

That can be handled like this:

$ref: usb-switch.yaml
properties:
  orientation-switch: true
additionalProperties: false

Though if you need unevaluatedProperties for some other reason, then 
that won't enforce it and it's just documentation. In that case, then 
perhaps usb-switch.yaml is not the right granularity and should be split 
up.

I put this into the category of "this is the least of our problems". I'm 
not that interested in enforcing what common properties a device uses or 
not. It's undocumented properties I'm worried about or lack of 
constraints (on reg, clocks, interrupts, etc.).

Rob