Re: [EXT] Re: [PATCH 1/2] dt-bindings: watchdog: marvell octeonTX2 GTI system atchdog driver

[Guenter Roeck]

On Fri, Apr 14, 2023 at 11:34:36AM +0000, Bharat Bhushan wrote:
> Please see inline 
> 
> > -----Original Message-----
> > From: Krzysztof Kozlowski <krzysztof.kozlowski@xxxxxxxxxx>
> > Sent: Friday, April 14, 2023 5:02 PM
> > To: Bharat Bhushan <bbhushan2@xxxxxxxxxxx>; wim@xxxxxxxxxxxxxxxxxx;
> > linux@xxxxxxxxxxxx; robh+dt@xxxxxxxxxx; krzysztof.kozlowski+dt@xxxxxxxxxx;
> > linux-watchdog@xxxxxxxxxxxxxxx; devicetree@xxxxxxxxxxxxxxx; linux-
> > kernel@xxxxxxxxxxxxxxx
> > Subject: Re: [EXT] Re: [PATCH 1/2] dt-bindings: watchdog: marvell octeonTX2 GTI
> > system atchdog driver
> > 
> > On 14/04/2023 13:29, Bharat Bhushan wrote:
> > 
> > >
> > >>
> > >>> +
> > >>> +examples:
> > >>> +  - |
> > >>> +    soc {
> > >>> +        #address-cells = <2>;
> > >>> +        #size-cells = <2>;
> > >>> +
> > >>> +        watchdog@802000040000 {
> > >>> +          compatible = "marvell-octeontx2-wdt";
> > >>> +          reg = <0x8020 0x40000 0x0 0x20000>;
> > >>
> > >> Are you sure that this is correct DTS? 32-bit numbers are usually
> > >> 8-digit long. Plus size of 0x20000 is crazy huge. And the unit
> > >> address is a bit unusual. Are you sure dtc W=1 does not say about any errors in
> > your DTS?
> > >
> > > Each cell is 32bit, so if we specify less than upper values becomes zeros (0s).
> > 
> > ... and what is the convention/coding style of your subarch? Is to have short or 0-
> > padded reg addresses?
> > 
> > Anyway, I have doubts this was tested, so please confirm that dtbs W=1 and
> > dtbs_check produce no errors on your DTS. BTW, where is the DTS?
> 
> Our device tree are part of firmware and not in Linux.

Odd argument. That is true for pretty much every devicetree file.

Guenter