On Tue, 09 Jun 2020, Rob Herring wrote:

On Tue, Jun 09, 2020 at 05:01:17PM +0200, Michael Walle wrote:

quoted

Am 2020-06-09 16:42, schrieb Mark Brown:

quoted

On Tue, Jun 09, 2020 at 04:38:31PM +0200, Michael Walle wrote:

quoted

  mfd-device@10 {
    compatible = "simple-regmap", "simple-mfd";
    reg = <10>;
    regmap,reg-bits = <8>;
    regmap,val-bits = <8>;
    sub-device@0 {
      compatible = "vendor,sub-device0";
      reg = <0>;
    };

A DT binding like this is not a good idea, encoding the details of the
register map into the DT binding makes it an ABI which is begging for
trouble.  I'd also suggest that any device using a generic driver like
this should have a specific compatible string for the device so we can
go back and add quirks later if we need them.

Like in the spidev case, yes. But OTOH if I _just_ encode the parameters
for the regmap a MFD, Lee don't agree because its just a shim. So either
way I seem to be stuck here.

Where should I put the code to create an i2c driver, init a regmap and
populate its childen?

Find another driver doing this already and rename it 'simple-mfd' (no 
relation to the DT binding) and add your compatible string to it. 
'Generic' or 'simple' drivers don't require generic/simple DT bindings.

Creating a generic driver is one of the options spinning around in my
head.  If nothing better comes of these discussions, I'll turn my hand
to it soon.

Or extend the existing syscon driver to look up the bus_type and create 
the regmap based on the bus type?

-- 
Lee Jones [李琼斯]
Senior Technical Lead - Developer Services
Linaro.org │ Open source software for Arm SoCs
Follow Linaro: Facebook | Twitter | Blog

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel