[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <0ef4d4ee-f8b7-4fca-b382-ebaa6b7ee050@kernel.org>
Date: Tue, 13 Aug 2024 18:13:24 +0200
From: Krzysztof Kozlowski <krzk@...nel.org>
To: Alexander Stein <alexander.stein@...tq-group.com>,
Markus Niebel <Markus.Niebel@...tq-group.com>, 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>, Fabio Estevam <festevam@...il.com>
Cc: Pengutronix Kernel Team <kernel@...gutronix.de>,
devicetree@...r.kernel.org, imx@...ts.linux.dev,
linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
linux@...tq-group.com
Subject: Re: [PATCH 5/5] ARM: dts: imx6qdl: Rename USB hub node name
On 13/08/2024 13:02, Alexander Stein wrote:
> Am Dienstag, 13. August 2024, 11:44:28 CEST schrieb Krzysztof Kozlowski:
>> On 13/08/2024 11:27, Alexander Stein wrote:
>>> Am Dienstag, 13. August 2024, 11:20:08 CEST schrieb Krzysztof Kozlowski:
>>>> On 12/08/2024 16:34, Markus Niebel wrote:
>>>>> From: Alexander Stein <alexander.stein@...tq-group.com>
>>>>>
>>>>> According to microchip,usb2514.yaml the node name shall be usb-hub.
>>>>
>>>> That's not true. The schema does not say anything like this. Old name is
>>>> correct. NAK.
>>>
>>> So, is the schema incorrect? There is the dtbs_check warning:
>>> arch/arm/boot/dts/nxp/imx/imx6q-mba6b.dtb: hub@1: $nodename:0: 'hub@1' does not match '^usb(@.*)?'
>>> from schema $id: http://devicetree.org/schemas/usb/microchip,usb2514.yaml#
>>
>> If you have a warning, shorten it and paste it so this will be obvious.
>> If you look at several bindings, the hub is widely used name. I think
>> the schema is not correct here - I do not see any properties from
>> usb.yaml being used here (for usb2514). What's more, if you compare
>> usb2514 with any other on-board HUB representations (because that's the
>> only point why we have it in bindings, right?), none of them reference
>> usb(-hcd)?.yaml.
>>
>> These are not USB controllers, IMO.
>
> I raised that concern in [1] already, but nobody commented.
With 200 new emails in my inbox every day, I bet I miss many threads...
The nodes could be named usb-hub, makes sense, but:
1. Several bindings examples have hub, so we would need to correct all
of them for consistency?
2. For me hub is generic and obvious enough. Just like "phy", even
though DT spec has "usb-phy".
3. Renaming nodes to fix issues is very good. Renaming just to make
things more readable is also good or because coding style asks for it.
Renaming just because one person likes hub, other likes usb-hub, so just
personal preference, is for me churn. Considering it sometimes breaks
some users (it is not ABI... till it is), personal preference is poor
reason.
I think the schema should be fixed, because this is not an USB
controller, in terms how we call things which are "USB controllers". On
the bus it is USB device and USB hub. For USB device we have schema. For
USB hub we don't, so it does not matter.
Best regards,
Krzysztof
Powered by blists - more mailing lists