Re: [PATCH 1/3] dt-bindings: net: can: m_can: Add Documentation for stb-gpios

From: Faiz Abbas <faiz_abbas@ti.com>
To: Rob Herring <robh@kernel.org>
Cc: Dan Murphy <dmurphy@ti.com>, Sekhar Nori <nsekhar@ti.com>,
	<linux-arm-kernel@lists.infradead.org>,
	<linux-kernel@vger.kernel.org>, <devicetree@vger.kernel.org>,
	<netdev@vger.kernel.org>, <linux-can@vger.kernel.org>,
	<catalin.marinas@arm.com>, <mark.rutland@arm.com>,
	<davem@davemloft.net>, <mkl@pengutronix.de>, <wg@grandegger.com>,
	<sriram.dash@samsung.com>, <nm@ti.com>, <t-kristo@ti.com>
Subject: Re: [PATCH 1/3] dt-bindings: net: can: m_can: Add Documentation for stb-gpios
Date: Mon, 17 Feb 2020 19:23:52 +0530	[thread overview]
Message-ID: <15ae4d0e-10cf-7b4b-6487-8b64f885f184@ti.com> (raw)
In-Reply-To: <20200203120610.GA9303@bogus>

Rob,

On 03/02/20 5:36 pm, Rob Herring wrote:
> On Thu, Jan 23, 2020 at 01:09:41PM +0530, Faiz Abbas wrote:
>> Hi,
>>
>> On 22/01/20 8:04 pm, Dan Murphy wrote:
>>> Sekhar
>>>
>>> On 1/22/20 8:24 AM, Sekhar Nori wrote:
>>>> On 22/01/20 7:05 PM, Dan Murphy wrote:
>>>>> Faiz
>>>>>
>>>>> On 1/22/20 2:03 AM, Faiz Abbas wrote:
>>>>>> The CAN transceiver on some boards has an STB pin which is
>>>>>> used to control its standby mode. Add an optional property
>>>>>> stb-gpios to toggle the same.
>>>>>>
>>>>>> Signed-off-by: Faiz Abbas <faiz_abbas@ti.com>
>>>>>> Signed-off-by: Sekhar Nori <nsekhar@ti.com>
>>>>>> ---
>>>>>>    Documentation/devicetree/bindings/net/can/m_can.txt | 2 ++
>>>>>>    1 file changed, 2 insertions(+)
>>>>>>
>>>>>> diff --git a/Documentation/devicetree/bindings/net/can/m_can.txt
>>>>>> b/Documentation/devicetree/bindings/net/can/m_can.txt
>>>>>> index ed614383af9c..cc8ba3f7a2aa 100644
>>>>>> --- a/Documentation/devicetree/bindings/net/can/m_can.txt
>>>>>> +++ b/Documentation/devicetree/bindings/net/can/m_can.txt
>>>>>> @@ -48,6 +48,8 @@ Optional Subnode:
>>>>>>                  that can be used for CAN/CAN-FD modes. See
>>>>>>                
>>>>>> Documentation/devicetree/bindings/net/can/can-transceiver.txt
>>>>>>                  for details.
>>>>>> +stb-gpios        : gpio node to toggle the STB (standby) signal on
>>>>>> the transceiver
>>>>>> +
>>>>> The m_can.txt is for the m_can framework.  If this is specific to the
>>>>> platform then it really does not belong here.
>>>>>
>>>>> If the platform has specific nodes then maybe we need a
>>>>> m_can_platform.txt binding for specific platform nodes.  But I leave
>>>>> that decision to Rob.
>>>> Since this is transceiver enable, should this not be in
>>>> Documentation/devicetree/bindings/net/can/can-transceiver.txt?
>>>
>>
>> The transceiver node is just a node without an associated device. I had
>> tried to convert it to a phy implementation but that idea got shot down
>> here:
>>
>> https://lore.kernel.org/patchwork/patch/1006238/
> 
> Nodes and drivers are not a 1-1 thing. Is the transceiver a separate h/w 
> device? If so, then it should be a separate node and properties of that 
> device go in its node. 

The transceiver is indeed a separate device.

Also, nothing is stopping you from using the PHY
> binding without using the kernel's PHY framework.

The phy framework seemed like the best code reuse to implement it.

> 
> As to whether it should be a separate phy driver, I think probably the 
> wrong decision was made. We always seem to start out with no PHY on 
> these things and the complexity just grows until we need one. 
> 

We should be able to handle two properties (one max-datarate and the
other regulator node) for now. If we have to add more complex parts then
maybe we can think about the driver. I am just adding a xceiver
regulator for now.

Thanks,
Faiz