Re: [PATCH] dt-bindings: iio: admv1014: make all regs required

[Jonathan Cameron <jic23@kernel.org>]

On Thu, 27 Jul 2023 12:02:04 +0000
"Miclaus, Antoniu" <Antoniu.Miclaus@analog.com> wrote:

> > -----Original Message-----
> > From: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
> > Sent: Thursday, July 27, 2023 2:47 PM
> > To: Miclaus, Antoniu <Antoniu.Miclaus@analog.com>; jic23@kernel.org;
> > robh+dt@kernel.org; krzysztof.kozlowski+dt@linaro.org;
> > conor+dt@kernel.org; linux-iio@vger.kernel.org;
> > devicetree@vger.kernel.org; linux-kernel@vger.kernel.org
> > Subject: Re: [PATCH] dt-bindings: iio: admv1014: make all regs required
> > 
> > [External]
> > 
> > On 27/07/2023 13:31, Antoniu Miclaus wrote:  
> > > Since the regulators are required in the driver implementation, make
> > > them required also in the bindings.

This bit is probably not strictly true.  Try no providing them and you will
probably find stub regulators created on assumption they are fixed regs
that we aren't interested in controlling.

> > >  
> > 
> > The true reason should be whether the hardware requires them. Because if
> > hardware does not need some, the driver should be fixed.  
> The datasheet is not very explicit on this topic, but all the specifications of the
> part are built around these pins being supplied.
> Moreover, the evaluation board or the part comes with all VCC pins connected
> to fixed supplies.

Agreed - these should be required in the binding.

There was an old disagreement on this in which it they were not marked
required in some bindings because the regulator subsystem would assume they
were fixed supplies that were just missing in the DT and so provides stub regulators.
My understanding at least has changed and now we mark the required even if
the driver works fine on some boards without them being supplied.

Note though that this means the whole of trivial-devices.yaml is garbage
as very few devices actually work without any power :)

Jonathan



> > Best regards,
> > Krzysztof  
>