| Message ID | 20240319-b4-qcom-livetree-v3-12-e1b38d9b4fa4@linaro.org |
|---|---|
| State | New |
| Headers | show |
| Series | Qualcomm platform USB support | expand |
On Tue, 19 Mar 2024 at 17:52, Caleb Connolly <caleb.connolly@linaro.org> wrote: > > The USB VBUS supply for the type-A port is enabled via a GPIO regulator. > This is incorrectly modelled in Linux where only the PCIe dependency is > expressed. Can we send a fix for the Linux kernel DTS to correctly model it? We can then later get rid of this modification once that is accepted. -Sumit > Add a U-Boot specific dtsi snippet so that this supply will > get enabled when initialising USB. > > Signed-off-by: Caleb Connolly <caleb.connolly@linaro.org> > --- > arch/arm/dts/sdm845-db845c-u-boot.dtsi | 8 ++++++++ > 1 file changed, 8 insertions(+) > > diff --git a/arch/arm/dts/sdm845-db845c-u-boot.dtsi b/arch/arm/dts/sdm845-db845c-u-boot.dtsi > new file mode 100644 > index 000000000000..3c83e21251c4 > --- /dev/null > +++ b/arch/arm/dts/sdm845-db845c-u-boot.dtsi > @@ -0,0 +1,8 @@ > +// SPDX-License-Identifier: GPL-2.0 > + > +/* This GPIO must be turned on to enable the 5v VBUS > + * supply on the USB port. > + */ > +&usb_2_dwc3 { > + vbus-supply = <&pcie0_3p3v_dual>; > +}; > > -- > 2.44.0 >
On 19/03/2024 13:55, Sumit Garg wrote: > On Tue, 19 Mar 2024 at 17:52, Caleb Connolly <caleb.connolly@linaro.org> wrote: >> >> The USB VBUS supply for the type-A port is enabled via a GPIO regulator. >> This is incorrectly modelled in Linux where only the PCIe dependency is >> expressed. > > Can we send a fix for the Linux kernel DTS to correctly model it? We > can then later get rid of this modification once that is accepted. I spoke to Bjorn about this and apparently the correct way to model this will be to have a usb-connector node with a vbus-supply property. There is some work underway in Linux to support this kind of thing already. In the mean time he suggested to just make the regulator always-on, so I sent a patch to do that [1]. I actually hit this issue when booting from USB as the PCIe drivers aren't available in the initramfs, so USB never turns on, and the always-on hack fixes that. In the mean time, we'll still need this vbus-supply reference as U-Boot doesn't automatically probe regulators. > > -Sumit > >> Add a U-Boot specific dtsi snippet so that this supply will >> get enabled when initialising USB. >> >> Signed-off-by: Caleb Connolly <caleb.connolly@linaro.org> >> --- >> arch/arm/dts/sdm845-db845c-u-boot.dtsi | 8 ++++++++ >> 1 file changed, 8 insertions(+) >> >> diff --git a/arch/arm/dts/sdm845-db845c-u-boot.dtsi b/arch/arm/dts/sdm845-db845c-u-boot.dtsi >> new file mode 100644 >> index 000000000000..3c83e21251c4 >> --- /dev/null >> +++ b/arch/arm/dts/sdm845-db845c-u-boot.dtsi >> @@ -0,0 +1,8 @@ >> +// SPDX-License-Identifier: GPL-2.0 >> + >> +/* This GPIO must be turned on to enable the 5v VBUS >> + * supply on the USB port. >> + */ >> +&usb_2_dwc3 { >> + vbus-supply = <&pcie0_3p3v_dual>; >> +}; >> >> -- >> 2.44.0 >>
On 20/03/2024 12:33, Caleb Connolly wrote: > > > On 19/03/2024 13:55, Sumit Garg wrote: >> On Tue, 19 Mar 2024 at 17:52, Caleb Connolly <caleb.connolly@linaro.org> wrote: >>> >>> The USB VBUS supply for the type-A port is enabled via a GPIO regulator. >>> This is incorrectly modelled in Linux where only the PCIe dependency is >>> expressed. >> >> Can we send a fix for the Linux kernel DTS to correctly model it? We >> can then later get rid of this modification once that is accepted. > > I spoke to Bjorn about this and apparently the correct way to model this > will be to have a usb-connector node with a vbus-supply property. There > is some work underway in Linux to support this kind of thing already. > > In the mean time he suggested to just make the regulator always-on, so I > sent a patch to do that [1]. I actually hit this issue when booting from > USB as the PCIe drivers aren't available in the initramfs, so USB never > turns on, and the always-on hack fixes that. > > In the mean time, we'll still need this vbus-supply reference as U-Boot > doesn't automatically probe regulators. oh, forgot the link! [1]: https://lore.kernel.org/linux-arm-msm/20240320122515.3243711-1-caleb.connolly@linaro.org/ >> >> -Sumit >> >>> Add a U-Boot specific dtsi snippet so that this supply will >>> get enabled when initialising USB. >>> >>> Signed-off-by: Caleb Connolly <caleb.connolly@linaro.org> >>> --- >>> arch/arm/dts/sdm845-db845c-u-boot.dtsi | 8 ++++++++ >>> 1 file changed, 8 insertions(+) >>> >>> diff --git a/arch/arm/dts/sdm845-db845c-u-boot.dtsi b/arch/arm/dts/sdm845-db845c-u-boot.dtsi >>> new file mode 100644 >>> index 000000000000..3c83e21251c4 >>> --- /dev/null >>> +++ b/arch/arm/dts/sdm845-db845c-u-boot.dtsi >>> @@ -0,0 +1,8 @@ >>> +// SPDX-License-Identifier: GPL-2.0 >>> + >>> +/* This GPIO must be turned on to enable the 5v VBUS >>> + * supply on the USB port. >>> + */ >>> +&usb_2_dwc3 { >>> + vbus-supply = <&pcie0_3p3v_dual>; >>> +}; >>> >>> -- >>> 2.44.0 >>> >
On Wed, 20 Mar 2024 at 18:04, Caleb Connolly <caleb.connolly@linaro.org> wrote: > > > > On 20/03/2024 12:33, Caleb Connolly wrote: > > > > > > On 19/03/2024 13:55, Sumit Garg wrote: > >> On Tue, 19 Mar 2024 at 17:52, Caleb Connolly <caleb.connolly@linaro.org> wrote: > >>> > >>> The USB VBUS supply for the type-A port is enabled via a GPIO regulator. > >>> This is incorrectly modelled in Linux where only the PCIe dependency is > >>> expressed. > >> > >> Can we send a fix for the Linux kernel DTS to correctly model it? We > >> can then later get rid of this modification once that is accepted. > > > > I spoke to Bjorn about this and apparently the correct way to model this > > will be to have a usb-connector node with a vbus-supply property. There > > is some work underway in Linux to support this kind of thing already. > > > > In the mean time he suggested to just make the regulator always-on, so I > > sent a patch to do that [1]. I actually hit this issue when booting from > > USB as the PCIe drivers aren't available in the initramfs, so USB never > > turns on, and the always-on hack fixes that. > > > > In the mean time, we'll still need this vbus-supply reference as U-Boot > > doesn't automatically probe regulators. > Try adding regulators_enable_boot_on() to the common board code. With that there shouldn't be any need for this vbus-supply reference. > oh, forgot the link! > > [1]: > https://lore.kernel.org/linux-arm-msm/20240320122515.3243711-1-caleb.connolly@linaro.org/ -Sumit
diff --git a/arch/arm/dts/sdm845-db845c-u-boot.dtsi b/arch/arm/dts/sdm845-db845c-u-boot.dtsi new file mode 100644 index 000000000000..3c83e21251c4 --- /dev/null +++ b/arch/arm/dts/sdm845-db845c-u-boot.dtsi @@ -0,0 +1,8 @@ +// SPDX-License-Identifier: GPL-2.0 + +/* This GPIO must be turned on to enable the 5v VBUS + * supply on the USB port. + */ +&usb_2_dwc3 { + vbus-supply = <&pcie0_3p3v_dual>; +};
The USB VBUS supply for the type-A port is enabled via a GPIO regulator. This is incorrectly modelled in Linux where only the PCIe dependency is expressed. Add a U-Boot specific dtsi snippet so that this supply will get enabled when initialising USB. Signed-off-by: Caleb Connolly <caleb.connolly@linaro.org> --- arch/arm/dts/sdm845-db845c-u-boot.dtsi | 8 ++++++++ 1 file changed, 8 insertions(+)