Hi Simon,
Simon Glass <[email protected]> wrote:
Well, yes, mt7981_pinctrl is wrong since it is not actually binding
the GPIO devices until it itself is probed. It should do it when it is
bound.
Oh I see! Yes, I can see the mtk_gpiochip_register(dev) in
mtk_pinctrl_common_probe() exactly as you say.
Better still, those GPIO devices should be in the devicetree and bound
automatically by driver model. But, sigh, I see that there is no
compatible string in the gpio subnode of pinctrl@11d00000. It should
really have one and avoid all this pointless code and problems.
mtk_pinctrl_common_probe() is misnamed, as it actually binds and then
probes.
So (unless Linux allows a patch to add a compatible string) it needs a
new mtk_pinctrl_common_bind() (called from mtk_pinctrl_mt7981_bind())
which calls mtk_gpiochip_register(). Then you won't need to add your
dev_or_flags() into mtk_pinctrl_mt7981_bind().
Yes, that makes complete sense. Many thanks! I'm very happy to write that
patch and grab back the physical hardware to double-check on if you like?
(Or equally happy to leave it if you'd prefer to fix yourself?)
Presumably it needs to apply to every mtk soc that uses
mtk_pinctrl_common_probe() as they'll all be affected by this problem.
Best wishes,
Chris.
