On Tue, Jul 2, 2024 at 1:32 PM Dmitry Baryshkov
<dmitry.baryshkov@xxxxxxxxxx> wrote:
>
> On Tue, Jul 02, 2024 at 01:13:09PM GMT, Bartosz Golaszewski wrote:
> > On Tue, Jul 2, 2024 at 1:10 PM Dmitry Baryshkov
> > <dmitry.baryshkov@xxxxxxxxxx> wrote:
> > >
> > > On Tue, Jul 02, 2024 at 12:42:02PM GMT, Krzysztof Kozlowski wrote:
> > > > On 02/07/2024 11:16, Amit Pundir wrote:
> > > > > Describe the ath12k WLAN on-board the WCN7850 module present on the
> > > > > board.
> > > > >
> > > > > Signed-off-by: Amit Pundir <amit.pundir@xxxxxxxxxx>
> > > > > ---
> > > > > Kanged verbatim from 490812872449 ("arm64: dts: qcom: sm8550-qrd: add the Wifi node").
> > > > >
> > > > > arch/arm64/boot/dts/qcom/sm8550-hdk.dts | 97 +++++++++++++++++++++++++
> > > > > 1 file changed, 97 insertions(+)
> > > > >
> > > > > diff --git a/arch/arm64/boot/dts/qcom/sm8550-hdk.dts b/arch/arm64/boot/dts/qcom/sm8550-hdk.dts
> > > > > index 12d60a0ee095..c453d081a2df 100644
> > > > > --- a/arch/arm64/boot/dts/qcom/sm8550-hdk.dts
> > > > > +++ b/arch/arm64/boot/dts/qcom/sm8550-hdk.dts
> > > > > @@ -279,6 +279,68 @@ platform {
> > > > > };
> > > > > };
> > > > > };
> > > > > +
> > > > > + wcn7850-pmu {
> > > > > + compatible = "qcom,wcn7850-pmu";
> > > > > +
> > > > > + pinctrl-names = "default";
> > > > > + pinctrl-0 = <&wlan_en>, <&pmk8550_sleep_clk>;
> > > > > +
> > > > > + wlan-enable-gpios = <&tlmm 80 GPIO_ACTIVE_HIGH>;
> > > > > + /*
> > > > > + * TODO Add bt-enable-gpios once the Bluetooth driver is
> > > > > + * converted to using the power sequencer.
> > > >
> > > > I don't understand why hardware description should depend on the driver.
> > > > Either you have this GPIO or not. If you have it, what does it matter if
> > > > there is no driver who can play with it?
> > >
> > > Then there is a conflict between BT and PMU, which both will try to
> > > access the gpio (or at least the pinctrl).
> >
> > Ah, so this slipped through the cracks. Amit merely copied it from existing dts.
> >
> > Yes, there's a conflict but Krzysztof is right - we should handle this
> > in the driver, not in dts.
> >
> > However for that we need a patch for the PMU pwrseq driver first. Let
> > me cook something up.
>
> Or for the BT driver, which might be more futureproof.
The problem here is that we have DT bindings that define the Bluetooth
node properties (including the bt-enable GPIO) for wcn7850 and so are
committed to supporting existing device-trees. Any change to the
handling of this model (unlike QCA6390) must retain backward
compatibility. I don't know yet how to approach it so for now I
propose to add a quirk to the pwrseq driver and revisit later.
Bart
