[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <b7fdd6d7-fd56-4e3a-84da-191411f9edd6@kernel.org>
Date: Thu, 18 Dec 2025 11:12:40 +0100
From: Krzysztof Kozlowski <krzk@...nel.org>
To: Arun Muthusamy <arun.muthusamy@...sler.com>
Cc: robh@...nel.org, krzk+dt@...nel.org, conor+dt@...nel.org,
mkl@...gutronix.de, mailhol@...nel.org, devicetree@...r.kernel.org,
linux-kernel@...r.kernel.org, linux-can@...r.kernel.org
Subject: Re: [PATCH 02/10] dt-bindings: net: can: grcan: Convert GRCAN CAN
controllers binding from txt to YAML
On 11/12/2025 11:11, Arun Muthusamy wrote:
> Hi Krzysztof,
>
> Thank you for your thorough review. I’d like to clarify a few points
> regarding the DT binding and get your guidance.
>
> On 11/18/25 12:01, Krzysztof Kozlowski wrote:
>> On 18/11/2025 10:21, Arun Muthusamy wrote:
>>> + Fallback on node name matching for systems that don't provide compatible.
>>> + enum:
>>> + - GAISLER_GRCAN
>>> + - 01_03d
>>> + - GAISLER_GRHCAN
>>> + - "01_034"
>> This does not really work. Are you really defining here "name" property?
>
> The driver supports two of the platforms which are LEON and NOEL
I don't care much about driver. Please describe here hardware.
> platforms. PROM-based *LEON* systems identify uses the "node name"
> property, while DTS based *NOEL* systems use proper "|compatible"|strings. On LEON (SPARC32), AMBA Plug & Play information creates the DT
> properties, and drivers historically match devices based on node names.
> To reflect this, I updated the $nodename pattern to support LEON-style
> node names: properties:
> $nodename:
> pattern: "^(GAISLER_GRCAN|01_03d|GAISLER_GRHCAN|01_034)$" I’d
> appreciate any suggestions on the preferred way to describe this node
> name for PROM-based LEON.
>
>>> +
>>> + reg:
>>> + maxItems: 1
>>> +
>>> + interrupts:
>>> + maxItems: 1
>>> +
>>> + clocks:
>>> + maxItems: 1
>>> +
>>> + freq:
>>> + $ref: /schemas/types.yaml#/definitions/uint32
>>> + description: |
>>> + Frequency of the external oscillator clock in Hz (the frequency of the
>>> + amba bus in the ordinary case).
>>> + This property should be used by systems that utilize the common clock
>>> + framework is not supported.
>> Missing systemid. Your commit msg must explain any changes done to the
>> binding during conversion.
>
> The driver now reads systemid directly from /ambapp0, so the property no
> longer needs to be defined in the DTS. The previous documentation was
> outdated and should have been updated after commit:
> 1e93ed26acf0 ("can: grcan: grcan_probe(): fix broken system id check for
> errata workaround needs").
Read again what I wrote.
>
>>> + - reg
>>> + - interrupts
>> Where is freq? It was required in the old binding. Again, you need to
>> explain the changes.
>
> LEON: relies on the freq property
> NOEL: uses a standard clocks binding
> Because of this dual approach, the freq property is no longer required
> in the DTS binding itself as theAMBA Plug & Play creates the DT properties.
Please read my comment carefully: you need to explain the changes. But
not to me, to everyone in the commit msg.
Best regards,
Krzysztof
Powered by blists - more mailing lists