Re: [PATCH v2 1/3] dt-bindings: usb: qcom,snps-dwc3: Add mode-switch

[Joonhoe Kim]

On Wed, Nov 04, 2026 at 02:13:21PM +0100, Krzysztof Kozlowski wrote:
> On Tue, Nov 03, 2026 at 10:50:55PM +0900, Joonhoe Kim wrote:
> > The QMP USB3-DP combo PHY stops the USB3 PIPE clock when a Type-C
> > DisplayPort pin assignment gives it all four lanes for DisplayPort,
> > while the controller can keep running its USB 2.0 port. The controller
> > has to move to the UTMI clock before that happens, so it needs the
> > Type-C mux state as well. fwnode_typec_mux_get() only matches graph
>
> Controller does not perform the switch, does it?

No, it does not route any signals. It only needs to know the mode: in
pin assignments C and E it selects the UTMI clock as its PIPE clock
before the PHY stops the PIPE clock.

usb-switch.yaml describes mode-switch as "Possible handler of altmode
switching". The WCD939x codec uses it in a similar way: its mux
callback only reports the audio accessory mode to its headset
detection. The codec nodes in sm8650-qrd.dts and sm8750-mtp.dts have
mode-switch.

> > nodes that have the "mode-switch" property, so allow it on the
> > controller node.
>
> So judging by this, you would need to fix fwnode_typec_mux_get() or
> opther driver code, not the binding?

The binding is needed because typec_mux_match() checks this property
before the mux driver has probed (see the comment in typec_mux_match(),
drivers/usb/typec/mux.c). Without that check, any graph neighbour that
is not a mux would make fwnode_typec_mux_get() return -EPROBE_DEFER
forever.

Apple and Rockchip handle this in the PHY driver, because there the PHY
driver can reach the controller-side USB3 control (the Apple PIPE
handler, the Rockchip USB GRF). On Qualcomm the clock
select is in the dwc3 glue (QSCRATCH), so the alternative would be a new
interface for the QMP PHY to notify the controller. I found no existing
one. It would have one advantage: the PHY knows when it actually
switches. I can try that if you prefer it.

> > +  mode-switch: true
>
> Drop, but you might explicitly disallow orientation and retimer, if
> entire solution stays after review.

OK. v3 will keep the usb-switch.yaml reference, drop this line and set
orientation-switch and retimer-switch to false.

Thanks,
Joonhoe Kim