Re: [PATCH v21 08/13] mfd: core: Add firmware-node support to MFD cells

[Bartosz Golaszewski]

On Tue, Apr 28, 2026 at 2:18 PM Shivendra Pratap
<shivendra.pratap@xxxxxxxxxxxxxxxx> wrote:
>
>
>
> On 28-04-2026 13:33, Bartosz Golaszewski wrote:
> > On Mon, 27 Apr 2026 19:34:48 +0200, Shivendra Pratap
> > <shivendra.pratap@xxxxxxxxxxxxxxxx> said:
> >> MFD core has no way to register a child device using an explicit firmware
> >> node. This prevents drivers from registering child nodes when those nodes
> >> do not define a compatible string. One such example is the PSCI
> >> "reboot-mode" node, which omits a compatible string as it describes
> >> boot-states provided by the underlying firmware.
> >>
> >> Extend struct mfd_cell with an optional fwnode pointer and attach it to the
> >> child device during registration when no firmware node is already assigned.
> >>
> >> Signed-off-by: Shivendra Pratap <shivendra.pratap@xxxxxxxxxxxxxxxx>
> >> ---
> >
> > In the following patch you use this and violate the supposed constness of
> > MFD cells. This also mean you need to modify the mfd_cell at run-time. I would
> > prefer introducing a field in mfd_cell: bool use_parent_fwnode which would
> > indicate to MFD core that - in case no fwnode was assigned from other sources:
> > DT or software node - it should reuse the parent MFD device's node.
>
> Have one doubt on this, When we fallback to use_parent_fwnode for cell:
> psci-reboot-mode, at psci_reboot_mode_probe(), pDev->dev->of_node point
> to psci node.
>
> But we want dev->of_node to point to "reboot-mode" node. This may need
> an explicit assignment again of of_node?
>

Ah, sorry you're right. I'd like to avoid having to modify the
mfd_cell entry at run-time if possible.

How about a callback in mfd_cell like:

struct fwnode_handle *(*get_fwnode)(struct device *parent);

If set, it would be called from MFD core and would do the name lookup
and return the fwnode which would then be assigned to the cell device?

Bart