On Tue, Nov 6, 2018 at 9:05 PM Sven Van Asbroeck <thesven73@xxxxxxxxx> wrote:
> On Tue, Nov 6, 2018 at 1:31 PM Rob Herring <robh@xxxxxxxxxx> wrote:
>
> > If the host is not a h/w component, but just a s/w protocol then it
> > doesn't belong in DT. Perhaps it could be a library which the bridge
> > driver can call into.
>
> Anybus cards have an id register, which identifies what they are, so
> that the appropriate client driver may be instantiated.
> In that sense anybus is very suited to the bus/client abstraction of
> pci/usb/etc.
>
> > What are the resets connected to? The slots? Maybe you should model
> > the slots in DT.
> >
>
> Yes, the resets are ultimately connected to the slots.
> I'm happy to model the slots in DT. It makes sense, they are physical,
> hardware components.
It may also be necessary to have a way to add more DT properties to
a device in the slot that can not be probed by looking at the ID
register. Ideally it should not be needed, but we have added this
to most buses anyway: USB, PCI, MMC and others can all be
probed automatically for many devices, but there are some devices
that need a place where extra information can be stored, e.g.
an ethernet MAC address (in the absence of a PROM), an
external clock or reset line that is not part of the normal protocol,
or a more specific identification when two devices accidentally
have the same ID.
Arnd
