Re: [PATCH v4 2/3] arm: dts: nuvoton: npcm7xx: Drop redundant FIU clock-names

Andrew Jeffery <[email protected]> Wed, 10 Jun 2026 21:28:23 +0930
Newsgroups org.ozlabs.lists.openbmc,org.kernel.vger.linux-devicetree,org.kernel.vger.linux-kernel,org.kernel.vger.linux-spi
Message-ID <8a059e3a4a3338bc03829071a720012bd426bf6d.camel@codeconstruct.com.au>
On Wed, 2026-06-10 at 14:56 +0300, Tomer Maimon wrote:
> On Wed, 10 Jun 2026 at 13:49, Andrew Jeffery
> <[email protected]> wrote:
> > 
> > On Tue, 2026-06-09 at 19:39 +0300, Tomer Maimon wrote:
> > > The NPCM7xx FIU controller driver gets its single clock with
> > > devm_clk_get_enabled(dev, NULL) and does not perform a named
> > > clock lookup. Drop the redundant clock-names properties from the
> > > FIU controller nodes so the DTS describes only the resources the
> > > driver actually uses.
> > 
> > The devicetree is a description of the hardware in the form documented
> > by the bindings. Generally it's not right to discuss Linux drivers
> > here: they're only relevant in the context of Linux, but the devicetree
> > binding governs devicetrees over multiple projects.
> > 
> > From a quick look it seems that these names are not described in the
> > corresponding binding, therefore no drivers should be using them and as
> > such they can (and should) be dropped. A driver would only be worth
> > mentioning if it did use the undocumented names (as that would be a
> > complication).
> > 
> > Can you please rework the description?
> do you mean change the description as follows
> "
> The NPCM7xx FIU controller nodes reference a single clock,
> but the FIU binding does not describe their clock-names properties.
> Drop the undocumented names so the DTS matches the binding.
> "

That sounds reasonable to me.

Thanks,

Andrew