linux-kernel - Re: [PATCH 1/3] media: dt-bindings: nxp,imx8-isi: Add i.MX95 ISI compatible string

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: <20251103122320.GT27255@pendragon.ideasonboard.com>
Date: Mon, 3 Nov 2025 14:23:20 +0200
From: Laurent Pinchart <laurent.pinchart@...asonboard.com>
To: Frank Li <Frank.li@....com>
Cc: Guoniu Zhou <guoniu.zhou@....nxp.com>,
	Mauro Carvalho Chehab <mchehab@...nel.org>,
	Rob Herring <robh@...nel.org>,
	Krzysztof Kozlowski <krzk+dt@...nel.org>,
	Conor Dooley <conor+dt@...nel.org>, Shawn Guo <shawnguo@...nel.org>,
	Sascha Hauer <s.hauer@...gutronix.de>,
	Pengutronix Kernel Team <kernel@...gutronix.de>,
	Fabio Estevam <festevam@...il.com>, linux-media@...r.kernel.org,
	devicetree@...r.kernel.org, imx@...ts.linux.dev,
	linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
	Guoniu Zhou <guoniu.zhou@....com>
Subject: Re: [PATCH 1/3] media: dt-bindings: nxp,imx8-isi: Add i.MX95 ISI
 compatible string

On Mon, Oct 27, 2025 at 03:19:58PM -0400, Frank Li wrote:
> On Mon, Oct 27, 2025 at 12:04:38AM +0200, Laurent Pinchart wrote:
> > On Fri, Oct 24, 2025 at 09:34:14AM -0400, Frank Li wrote:
> > > On Fri, Oct 24, 2025 at 05:46:52PM +0800, Guoniu Zhou wrote:
> > > > From: Guoniu Zhou <guoniu.zhou@....com>
> > > >
> > > > The ISI module on i.MX95 supports up to eight channels and four link
> > > > sources to obtain the image data for processing in its pipelines. It
> > > > can process up to eight image sources at the same time.
> > > >
> > > > Signed-off-by: Guoniu Zhou <guoniu.zhou@....com>
> > > > ---
> > > >  .../devicetree/bindings/media/nxp,imx8-isi.yaml    | 26 +++++++++++++++++++++-
> > > >  1 file changed, 25 insertions(+), 1 deletion(-)
> > > >
> > > > diff --git a/Documentation/devicetree/bindings/media/nxp,imx8-isi.yaml b/Documentation/devicetree/bindings/media/nxp,imx8-isi.yaml
> > > > index f43b91984f0152fbbcf80db3b3bbad7e8ad6c11e..eaab98ecf343a2cd3620f7469c016c3955d37406 100644
> > > > --- a/Documentation/devicetree/bindings/media/nxp,imx8-isi.yaml
> > > > +++ b/Documentation/devicetree/bindings/media/nxp,imx8-isi.yaml
> > > > @@ -23,6 +23,7 @@ properties:
> > > >        - fsl,imx8mp-isi
> > > >        - fsl,imx8ulp-isi
> > > >        - fsl,imx93-isi
> > > > +      - fsl,imx95-isi
> > > >
> > > >    reg:
> > > >      maxItems: 1
> > > > @@ -49,7 +50,7 @@ properties:
> > > >    interrupts:
> > > >      description: Processing pipeline interrupts, one per pipeline
> > > >      minItems: 1
> > > > -    maxItems: 2
> > > > +    maxItems: 8
> > > >
> > > >    power-domains:
> > > >      maxItems: 1
> > > > @@ -109,6 +110,29 @@ allOf:
> > > >              - port@0
> > > >              - port@1
> > > >
> > > > +  - if:
> > > > +      properties:
> > > > +        compatible:
> > > > +          contains:
> > > > +            const: fsl,imx95-isi
> > > > +    then:
> > > > +      properties:
> > > > +        interrupts:
> > > > +          maxItems: 8
> > >
> > > should minItems: 8 because you already limit maxItems at top;
> >
> > As far as I understand, when no "items" are specified, minItems defaults
> > to 1, and maxItems defaults to minItems (if specified) or 0 (if minItems
> > is not specified). This is implemented in dtschema/lib.py of
> > https://github.com/devicetree-org/dt-schema.git.
> >
> > Then, in dtschema/fixups.py, if only one of minItems or maxItems is
> > specified, the other one is set to the same value. I believe relying on
> > this is frowned upon by the DT maintainers.
> >
> > We could specify minItems only here, as the top-level constraint will
> > ensure we don't go over 8. That's not very future-proof though, so I
> > think specifying both minItems and maxItems would be best. Confirmation
> > from a DT maintainer would be appreciated.
> 
> I pretty sure I am correct. please below thread, I met many similar cases
> before.
> https://lore.kernel.org/imx/72c29785-eb7a-4cc8-a74c-3aad50129a23@kernel.org/

I discussed this with Krzysztof on IRC last week. He said that in
conditional statements, both minItems and maxItems should be set, except
when one of them is a border constraint (being the same as the top-level
constraint). In that case it can be omitted.

So in this particular case you're right, we should specify minItems
here, not maxItems.

> > The fsl,imx8mp-isi block above should then be fixed. It currently only
> > has maxItems set, minItems should be set to 2 as well.

The fsl,imx8mp-isi conditional block should specify both minItems and
maxItems, and set both to 2.

> >
> > > > +        ports:
> > > > +          properties:
> > > > +            port@0:
> > > > +              description: Pixel Link Slave 0
> > > > +            port@1:
> > > > +              description: Pixel Link Slave 1
> > > > +            port@2:
> > > > +              description: MIPI CSI-2 RX 0
> > > > +            port@3:
> > > > +              description: MIPI CSI-2 RX 1
> > > > +          required:
> > > > +            - port@2
> > > > +            - port@3
> > > > +
> > >
> > >      else
> > >        properties:
> > >          interrupts:
> > >            maxItem: 2
> > >
> > > to keep the same restriction for existed compatible string.
> >
> > We already specify the number of interrupts in two separate conditional
> > blocks above, with any else statement (for all but fsl,imx8mp-isi first,
> > and then for fsl,imx8mp-isi). Both specify maxItems, so I think we're
> > fine.
> >
> > > >  additionalProperties: false
> > > >
> > > >  examples:

-- 
Regards,

Laurent Pinchart