[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <06f11844-2b85-469d-8729-a8757a566814@kernel.org>
Date: Thu, 19 Sep 2024 14:42:33 +0200
From: Krzysztof Kozlowski <krzk@...nel.org>
To: Christian Marangi <ansuelsmth@...il.com>
Cc: Wim Van Sebroeck <wim@...ux-watchdog.org>,
Guenter Roeck <linux@...ck-us.net>, Rob Herring <robh@...nel.org>,
Krzysztof Kozlowski <krzk+dt@...nel.org>, Conor Dooley
<conor+dt@...nel.org>, linux-watchdog@...r.kernel.org,
devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
Lorenzo Bianconi <lorenzo@...nel.org>, upstream@...oha.com
Subject: Re: [PATCH 1/2] dt-bindings: watchdog: airoha: document watchdog for
Airoha EN7581
On 19/09/2024 14:39, Christian Marangi wrote:
> On Thu, Sep 19, 2024 at 02:35:02PM +0200, Krzysztof Kozlowski wrote:
>> On 19/09/2024 14:26, Christian Marangi wrote:
>>> Document watchdog for Airoha EN7581. This SoC implement a simple
>>> watchdog that supports a max timeout of 28 seconds.
>>>
>>> The watchdog ticks on half the BUS clock and require the BUS frequency
>>> to be provided.
>>
>> Clock provider should implement clk_get_rate()...
>>
>
> The BUS clock is internal and not exposed to the system hence
> clk_get_rate is not possible saddly.
>
>>>
>>
>> ...
>>
>>> +maintainers:
>>> + - Christian Marangi <ansuelsmth@...il.com>
>>> +
>>> +allOf:
>>> + - $ref: watchdog.yaml#
>>> +
>>> +properties:
>>> + compatible:
>>> + const: airoha,en7581-wdt
>>> +
>>> + reg:
>>> + maxItems: 1
>>> +
>>> + clock-frequency:
>>> + description: BUS frequency in Hz (timer ticks at half the BUS freq)
>>> + const: 300000000
>>
>> Which bus frequency? Aren't you missing here clock input?
>
> I'm putting here property to describe the internal clock to what the
> watchdog is attached. Should I drop this and just hardcode it
> internally to the driver or maybe declare the clock to be 150000000
> directly?
If this stays, then please mention "internal watchdog bus frequency".
If this is internal and it is part of an SoC (so not board!) why would
we need it in DT? I would imagine this is fixed per SoC, thus deduced
from the compatible.
clock-frequency property is legacy and in general discouraged. This
might be an exception, but for that I would like to see more of
explanations.
>
> Tick frequency is already not well defined so I tought it was a good
> idea to describe it in DT.
>
Best regards,
Krzysztof
Powered by blists - more mailing lists