On Wed, Feb 23, 2022 at 06:59:27PM +0100, Clément Léger wrote:
> Mark Brown <broonie@kernel.org> a écrit :

> > This doesn't look like it's trying to use a DT on an ACPI system though?

> Ideally no, but it is a possibility mentionned by Andrew, use DT
> overlays on an ACPI system. This series did not took this way (yet).
> Andrew mentionned that it could potentially be done but judging by your
> comment, i'm not sure you agree with that.

That seems like it's opening a can of worms that might be best left
closed.

> > There's been some discussion on how to handle loadable descriptions for
> > things like FPGA but I don't recall it ever having got anywhere concrete
> > - I could have missed something.  Those are dynamic cases which are more
> > trouble though.  For something that's a PCI card it's not clear that we
> > can't just statically instanitate the devices from kernel code, that was
> > how the MFD subsystem started off although it's now primarily applied to
> > other applications.  That looks to be what's going on here?

> Yes, in this series, I used the MFD susbsytems with mfd_cells. These
> cells are attached with a swnode. Then, needed subsystems are
> modified to use the fwnode API to be able to use them with
> devices that have a swnode as a primary node.

Note that not all subsystems are going to be a good fit for fwnode, it's
concerning for the areas where ACPI and DT have substantially different
models like regulators.

> > There were separately some issues with people trying to create
> > completely swnode based enumeration mechanisms for things that required
> > totally independent code for handling swnodes which seemed very
> > concerning but it's not clear to me if that's what's going on here.

> The card is described entirely using swnode that in a MFD PCI
> driver, everything is described statically. The "enumeration" is static
> since all the devices are described in the driver and registered using
> mfd_add_device() at probe time. Thus, I don't think it adds an
> enumeration mechanism like you mention but I may be wrong.

This was all on the side parsing the swnodes rather than injecting the
data.