Re: [PATCH] dt-bindings: mark unimplemented AXI snps,kbbe snps,mb and snps,rb

[Conor Dooley <conor@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 1753 bytes --]

On Tue, Mar 10, 2026 at 09:48:25AM +0100, Krzysztof Kozlowski wrote:
> On 10/03/2026 09:43, Krzysztof Kozlowski wrote:
> > On 09/03/2026 18:49, Conor Dooley wrote:
> >> On Sat, Mar 07, 2026 at 12:05:18AM +0000, Russell King (Oracle) wrote:
> >>> Any chance of a review from the DT maintainers please?
> >>
> >> Sure.
> >>
> >>> On Wed, Mar 04, 2026 at 05:30:32PM +0000, Russell King (Oracle) wrote:
> >>>> Mark the AXI snps,kbbe snps,mb and snps,rb properties as not
> >>>> implemented. These were introduced by commit afea03656add ("stmmac:
> >>>> rework DMA bus setting and introduce new platform AXI structure").
> >>>> While stmmac has code to parse these properties and save their values,
> >>>> these are written to write-only struct members - no code in stmmac
> >>>> has ever read their value. Hence, these properties have been non-
> >>>> functional from day one.
> >>
> >> If they're not implemented and have never been used, could we delete
> 
> Maybe I misunderstood, did you mean delete from the drivers as well?

Correct. That's what I meant by "delete entirely".

> This would work as well, but we cannot only remove the dt-bindings.
> 
> 
> >> entirely? Note that they also appear down around line 600 in the file,
> >> in the stmmac-axi-config section.
> >>
> > 
> > We should not document what is and what is not implemented, thus this
> > patch is not correct either. Either you deprecate properties, based on
> > some argument, or you leave them untouched regardless whether they are
> > implemented or not. If someone re-implements them next week, are you
> > going to change the binding? Usually no, unless these are not really
> > hardware properties.
> 
> 
> 
> Best regards,
> Krzysztof

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]