Re: [PATCH] dt-bindings: PCI: Remove obsolete .txt docs

[Bjorn Helgaas <helgaas@kernel.org>]

On Wed, Apr 23, 2025 at 08:02:04AM -0500, Rob Herring wrote:
> On Tue, Apr 22, 2025 at 12:18 PM Bjorn Helgaas <helgaas@kernel.org> wrote:
> > On Fri, Apr 04, 2025 at 05:15:57PM -0500, Rob Herring (Arm) wrote:
> > > The content in these files has been moved to the schemas in dtschema.
> > > pci.txt is covered by pci-bus-common.yaml and pci-host-bridge.yaml.
> > > pci-iommu.txt is covered by pci-iommu.yaml. pci-msi.txt is covered in
> > > msi-map property in pci-host-bridge.yaml.
> >
> > I guess "dtschema" refers to
> > https://github.com/devicetree-org/dt-schema?
> 
> Yes.
> 
> > I kinda wish there was some direct link from the Linux kernel source
> > to dt-schema where all this information now lives (Requester ID
> > format, iommu-map, msi-map, linux,pci-domain, reg (and reference to
> > IEEE Std 1275-1994), interrupt mapping info, external-facing, etc).
> > Being a DT neophyte, I need all the help I can get ;)
> 
> Links to each property wouldn't really scale. Might as well copy all
> the common schemas into the kernel tree at that point.

Right.  In my ignorance, I found it useful to just grep the Linux
source for these properties, but that doesn't work any more.

> > There are a few dangling references to pci.txt:
> >
> >   Documentation/devicetree/bindings/pci/aardvark-pci.txt: - max-link-speed: see pci.txt
> >   Documentation/devicetree/bindings/pci/aardvark-pci.txt: - reset-gpios: see pci.txt
> >   Documentation/devicetree/bindings/pci/v3-v360epc-pci.txt:- bus-range: see pci.txt
> >   Documentation/devicetree/bindings/pci/v3-v360epc-pci.txt:  1275-1994 (see pci.txt) with the following restriction:
> 
> If no one cares enough to convert these, then I don't think the link matters.

I guess the descriptions of these now live at
https://github.com/devicetree-org/dt-schema/blob/main/dtschema/schemas/pci/pci-bus-common.yaml

There are a few Documentation/devicetree/bindings/ files that include
URLs like that, but maybe that's not considered a canonical source?

Bjorn