[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <mqea2nlysz4x3ff7xhg3fypgiyvrpqz6pwje4kavxoigrdlbtr@k6jw7jqsbkxr>
Date: Mon, 28 Aug 2023 15:41:35 +0300
From: Serge Semin <fancer.lancer@...il.com>
To: Krzysztof Kozlowski <krzysztof.kozlowski@...aro.org>
Cc: Keguang Zhang <keguang.zhang@...il.com>, netdev@...r.kernel.org,
devicetree@...r.kernel.org, linux-mips@...r.kernel.org,
linux-kernel@...r.kernel.org, Lee Jones <lee@...nel.org>,
Rob Herring <robh+dt@...nel.org>,
Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
Conor Dooley <conor+dt@...nel.org>,
"David S . Miller" <davem@...emloft.net>,
Eric Dumazet <edumazet@...gle.com>,
Jakub Kicinski <kuba@...nel.org>,
Paolo Abeni <pabeni@...hat.com>,
Thomas Bogendoerfer <tsbogend@...ha.franken.de>,
Giuseppe Cavallaro <peppe.cavallaro@...com>,
Alexandre Torgue <alexandre.torgue@...s.st.com>,
Jose Abreu <joabreu@...opsys.com>,
Serge Semin <Sergey.Semin@...kalelectronics.ru>
Subject: Re: [PATCH v3 2/4] dt-bindings: net: Add Loongson-1 Ethernet
Controller
On Mon, Aug 28, 2023 at 09:15:17AM +0200, Krzysztof Kozlowski wrote:
> On 27/08/2023 23:01, Serge Semin wrote:
> > Hi Krzysztof
> >
> > On Sun, Aug 27, 2023 at 09:56:06AM +0200, Krzysztof Kozlowski wrote:
> >> On 26/08/2023 23:04, Serge Semin wrote:
> >>>> + clock-names:
> >>>> + items:
> >>>> + - const: stmmaceth
> >>>
> >>> clock-names:
> >>> const: stmmaceth
> >>> ?
> >>
> >
> >> The existing syntax is correct. This is a string array.
> >
> > Could you please clarify whether it's a requirement (always specify
> > items: property for an array) or just an acceptable option (another
> > one is suggested in my comment)? I am asking because:
> > 1. In this case the "clock-names" array is supposed to have only one
> > item. Directly setting "const: stmmaceth" with no items: property
> > shall simplify it.
> > 2. There are single-entry "clock-names" property in the DT-bindings
> > defined as I suggested.
> > 3. There is a "compatible" property which is also a string array but
> > it can be defined as I suggested (omitting the items property).
> >
> > so based on all of that using the "items:"-based constraint here seems
> > redundant. Am I wrong to think like that? If so in what aspect?
>
> Syntax is correct in both cases. However the single list compatible
> *cannot grow*, while single list clock might, when developer notices
> that the binding was incomplete. People add binding matching drivers,
> not the hardware, thus having incomplete list of clocks is happening all
> the time.
So it's just a matter of maintainability. Got it. Thanks.
-Serge(y)
>
> Best regards,
> Krzysztof
>
Powered by blists - more mailing lists