Re: [PATCH RFC 3/7] dt-bindings: i3c: renesas,i3c: Add binding for Renesas I3C controller

From: Frank Li <Frank.li@nxp.com>
To: Rob Herring <robh@kernel.org>
Cc: Tommaso Merciai <tommaso.merciai.xr@bp.renesas.com>,
	Wolfram Sang <wsa+renesas@sang-engineering.com>,
	linux-renesas-soc@vger.kernel.org,
	Alexandre Belloni <alexandre.belloni@bootlin.com>,
	Krzysztof Kozlowski <krzk+dt@kernel.org>,
	Conor Dooley <conor+dt@kernel.org>,
	Geert Uytterhoeven <geert+renesas@glider.be>,
	Magnus Damm <magnus.damm@gmail.com>,
	linux-i3c@lists.infradead.org, devicetree@vger.kernel.org
Subject: Re: [PATCH RFC 3/7] dt-bindings: i3c: renesas,i3c: Add binding for Renesas I3C controller
Date: Wed, 25 Jun 2025 21:37:21 -0400	[thread overview]
Message-ID: <aFykUQR4ZOJ6pm3/@lizhi-Precision-Tower-5810> (raw)
In-Reply-To: <20250625200451.GA2117971-robh@kernel.org>

On Wed, Jun 25, 2025 at 03:04:51PM -0500, Rob Herring wrote:
> On Thu, Jun 12, 2025 at 04:51:42PM +0200, Tommaso Merciai wrote:
> > Hi Frank, Wolfram,
> >
> > Thanks both for your work.
> >
> > On 12/06/25 16:31, Wolfram Sang wrote:
> > > Hi Frank,
> > >
> > > > > +  clock-names:
> > > > > +    oneOf:
> > > > > +      - items:
> > > > > +          - const: pclk
> > > > > +          - const: tclk
> > > > > +      - items:
> > > > > +          - const: pclk
> > > > > +          - const: pclkrw
> > > > > +          - const: tclk
> > > >
> > > > Suggested clock name is, apb, trans, srfs.
>
> Suggested by who?

It is my suggestion base on
 - 1: needn't clk surfix. But you provide rule below.
 - 2: pclk, tclk ... can't provide any meansful information to user.
It is hard to know what's that without check binding doc. And most case
there are not description under clocks.

but some common clock-names term for IP, like apb, bus, which will be easy
to know apb (control access register). 'bus' feed into i3c bus clock speed.
'bus' is important for i3c to set 12Mhz or lower freqeuncy for old i2c
devices.

Frank

>
> > > > Needn't clk surfix and provide meanful name.
> > >
> > > ? Why can't we have the name from the datasheets? It will be
> > > unnecessarily hard for users to verify the correctness of the clock
> > > assignments if they need an artifical mapping of these names. If they
> > > are the same as in the datasheet, the same task becomes very easy.
> > >
> >
> > Agreed.
> > Since the name comes from the hw manual I think is better to leave it as is.
>
> Agreed.
>
> While 'clk' is considered redundant, naming clocks 'p' and 't' is too
> terse even for me. So the rule is clk is redundant unless it is with a
> single letter. The rules apply to s/w folks making up names, not h/w
> manuals.
>
> Rob
>
> --
> linux-i3c mailing list
> linux-i3c@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-i3c