Message ID | 20220119124327.v20.5.Ie0d2c1214b767bb5551dd4cad38398bd40e4466f@changeid (mailing list archive) |
---|---|
State | Superseded |
Headers | show |
Series | usb: misc: Add onboard_usb_hub driver | expand |
On Wed, Jan 19, 2022 at 12:43:45PM -0800, Matthias Kaehlcke wrote: > Add nodes for the onboard USB hub on trogdor devices. Remove the > 'always-on' property from the hub regulator, since the regulator > is now managed by the onboard_usb_hub driver. > > Signed-off-by: Matthias Kaehlcke <mka@chromium.org> > Reviewed-by: Stephen Boyd <swboyd@chromium.org> > Reviewed-by: Douglas Anderson <dianders@chromium.org> > --- No DT maintainer approval yet? :(
On Tue, Feb 08, 2022 at 11:57:20AM +0100, Greg Kroah-Hartman wrote: > On Wed, Jan 19, 2022 at 12:43:45PM -0800, Matthias Kaehlcke wrote: > > Add nodes for the onboard USB hub on trogdor devices. Remove the > > 'always-on' property from the hub regulator, since the regulator > > is now managed by the onboard_usb_hub driver. > > > > Signed-off-by: Matthias Kaehlcke <mka@chromium.org> > > Reviewed-by: Stephen Boyd <swboyd@chromium.org> > > Reviewed-by: Douglas Anderson <dianders@chromium.org> > > --- > > No DT maintainer approval yet? :( Bjorn usually just picks DT changes into the QCOM tree when they are ready, so I wouldn't interpret anything into the lack of an explicit Ack.
Hi, On Tue, Feb 8, 2022 at 11:21 AM Matthias Kaehlcke <mka@chromium.org> wrote: > > On Tue, Feb 08, 2022 at 11:57:20AM +0100, Greg Kroah-Hartman wrote: > > On Wed, Jan 19, 2022 at 12:43:45PM -0800, Matthias Kaehlcke wrote: > > > Add nodes for the onboard USB hub on trogdor devices. Remove the > > > 'always-on' property from the hub regulator, since the regulator > > > is now managed by the onboard_usb_hub driver. > > > > > > Signed-off-by: Matthias Kaehlcke <mka@chromium.org> > > > Reviewed-by: Stephen Boyd <swboyd@chromium.org> > > > Reviewed-by: Douglas Anderson <dianders@chromium.org> > > > --- > > > > No DT maintainer approval yet? :( > > Bjorn usually just picks DT changes into the QCOM tree when they are > ready, so I wouldn't interpret anything into the lack of an explicit > Ack. Right, so the expectation is that this patch wouldn't land through the USB tree but would instead land through the Qualcomm tree, probably a revision after the code lands in the USB tree to avoid dependency problems. -Doug
On Tue, Feb 15, 2022 at 09:54:54AM -0800, Doug Anderson wrote: > Hi, > > On Tue, Feb 8, 2022 at 11:21 AM Matthias Kaehlcke <mka@chromium.org> wrote: > > > > On Tue, Feb 08, 2022 at 11:57:20AM +0100, Greg Kroah-Hartman wrote: > > > On Wed, Jan 19, 2022 at 12:43:45PM -0800, Matthias Kaehlcke wrote: > > > > Add nodes for the onboard USB hub on trogdor devices. Remove the > > > > 'always-on' property from the hub regulator, since the regulator > > > > is now managed by the onboard_usb_hub driver. > > > > > > > > Signed-off-by: Matthias Kaehlcke <mka@chromium.org> > > > > Reviewed-by: Stephen Boyd <swboyd@chromium.org> > > > > Reviewed-by: Douglas Anderson <dianders@chromium.org> > > > > --- > > > > > > No DT maintainer approval yet? :( > > > > Bjorn usually just picks DT changes into the QCOM tree when they are > > ready, so I wouldn't interpret anything into the lack of an explicit > > Ack. > > Right, so the expectation is that this patch wouldn't land through the > USB tree but would instead land through the Qualcomm tree, probably a > revision after the code lands in the USB tree to avoid dependency > problems. But our tools pick up the whole series. I can't just do "i will pick patches 1-4 only" easily, and neither can any other maintainer. Why not just get their ack so that I know it can come through the USB tree? That's what normally happens for other changes like this where a driver change is required first. thanks, greg k-h
Hi, On Tue, Feb 15, 2022 at 10:55 AM Greg Kroah-Hartman <gregkh@linuxfoundation.org> wrote: > > On Tue, Feb 15, 2022 at 09:54:54AM -0800, Doug Anderson wrote: > > Hi, > > > > On Tue, Feb 8, 2022 at 11:21 AM Matthias Kaehlcke <mka@chromium.org> wrote: > > > > > > On Tue, Feb 08, 2022 at 11:57:20AM +0100, Greg Kroah-Hartman wrote: > > > > On Wed, Jan 19, 2022 at 12:43:45PM -0800, Matthias Kaehlcke wrote: > > > > > Add nodes for the onboard USB hub on trogdor devices. Remove the > > > > > 'always-on' property from the hub regulator, since the regulator > > > > > is now managed by the onboard_usb_hub driver. > > > > > > > > > > Signed-off-by: Matthias Kaehlcke <mka@chromium.org> > > > > > Reviewed-by: Stephen Boyd <swboyd@chromium.org> > > > > > Reviewed-by: Douglas Anderson <dianders@chromium.org> > > > > > --- > > > > > > > > No DT maintainer approval yet? :( > > > > > > Bjorn usually just picks DT changes into the QCOM tree when they are > > > ready, so I wouldn't interpret anything into the lack of an explicit > > > Ack. > > > > Right, so the expectation is that this patch wouldn't land through the > > USB tree but would instead land through the Qualcomm tree, probably a > > revision after the code lands in the USB tree to avoid dependency > > problems. > > But our tools pick up the whole series. I can't just do "i will pick > patches 1-4 only" easily, and neither can any other maintainer. > > Why not just get their ack so that I know it can come through the USB > tree? That's what normally happens for other changes like this where a > driver change is required first. Huh. That's the first time I've heard that and I'm pretty used to patches in a series going through different trees, but it would be OK w/ me if Bjorn was willing to Ack this. Bjorn: what say you? I guess alternatively Matthias could send two series: one with the code and a later one with the dts. -Doug
On 15/02/2022 19:55, Greg Kroah-Hartman wrote: > On Tue, Feb 15, 2022 at 09:54:54AM -0800, Doug Anderson wrote: >> Hi, >> >> On Tue, Feb 8, 2022 at 11:21 AM Matthias Kaehlcke <mka@chromium.org> wrote: >>> >>> On Tue, Feb 08, 2022 at 11:57:20AM +0100, Greg Kroah-Hartman wrote: >>>> On Wed, Jan 19, 2022 at 12:43:45PM -0800, Matthias Kaehlcke wrote: >>>>> Add nodes for the onboard USB hub on trogdor devices. Remove the >>>>> 'always-on' property from the hub regulator, since the regulator >>>>> is now managed by the onboard_usb_hub driver. >>>>> >>>>> Signed-off-by: Matthias Kaehlcke <mka@chromium.org> >>>>> Reviewed-by: Stephen Boyd <swboyd@chromium.org> >>>>> Reviewed-by: Douglas Anderson <dianders@chromium.org> >>>>> --- >>>> >>>> No DT maintainer approval yet? :( >>> >>> Bjorn usually just picks DT changes into the QCOM tree when they are >>> ready, so I wouldn't interpret anything into the lack of an explicit >>> Ack. >> >> Right, so the expectation is that this patch wouldn't land through the >> USB tree but would instead land through the Qualcomm tree, probably a >> revision after the code lands in the USB tree to avoid dependency >> problems. > > But our tools pick up the whole series. I can't just do "i will pick > patches 1-4 only" easily, and neither can any other maintainer. I don't have problems picking individual patches - either b4 am on each patch or on entire series and dropping later unneeded commits. > > Why not just get their ack so that I know it can come through the USB > tree? That's what normally happens for other changes like this where a > driver change is required first. DTS is a description of the hardware and we take it via separate branches of SoC-fami0ly repositories. These are always separated from the driver changes. Always. For several reasons: 1. By convention, 2. To be sure there is no dependency on driver code thus an ABI break, 3. To have a nice and clean history of DTS changes, properly organized. What is more, if this was coming via my Samsung SoC tree towards SoC folks, I could not take it in one branch. I would need to physically split it, otherwise Arnd/Olof would bounce back my pull request saying I am mixing DTS with driver. Of course you do not have such requirement - I am just saying that splitting DTS is quite common and proper way. Best regards, Krzysztof
On Tue 15 Feb 12:55 CST 2022, Greg Kroah-Hartman wrote: > On Tue, Feb 15, 2022 at 09:54:54AM -0800, Doug Anderson wrote: > > Hi, > > > > On Tue, Feb 8, 2022 at 11:21 AM Matthias Kaehlcke <mka@chromium.org> wrote: > > > > > > On Tue, Feb 08, 2022 at 11:57:20AM +0100, Greg Kroah-Hartman wrote: > > > > On Wed, Jan 19, 2022 at 12:43:45PM -0800, Matthias Kaehlcke wrote: > > > > > Add nodes for the onboard USB hub on trogdor devices. Remove the > > > > > 'always-on' property from the hub regulator, since the regulator > > > > > is now managed by the onboard_usb_hub driver. > > > > > > > > > > Signed-off-by: Matthias Kaehlcke <mka@chromium.org> > > > > > Reviewed-by: Stephen Boyd <swboyd@chromium.org> > > > > > Reviewed-by: Douglas Anderson <dianders@chromium.org> > > > > > --- > > > > > > > > No DT maintainer approval yet? :( > > > > > > Bjorn usually just picks DT changes into the QCOM tree when they are > > > ready, so I wouldn't interpret anything into the lack of an explicit > > > Ack. > > > > Right, so the expectation is that this patch wouldn't land through the > > USB tree but would instead land through the Qualcomm tree, probably a > > revision after the code lands in the USB tree to avoid dependency > > problems. > > But our tools pick up the whole series. I can't just do "i will pick > patches 1-4 only" easily, and neither can any other maintainer. > Most other maintainers uses -P to selectively pick the patches that applies to their subsystem. That said, I really do recognize the inconvenience on your part and the number of patches being sent your way. > Why not just get their ack so that I know it can come through the USB > tree? That's what normally happens for other changes like this where a > driver change is required first. > Because while the change looks good I don't think it's fine to take it through the USB tree - the dts tree typically looks like a shotgun hit across the dts files. And you and I have already seen several times that dts changes do conflict when you take some of them in the USB tree. Unfortunately I see only two ways around this problem, either you start picking selectively or I manage to convince all contributors that they must split their series to keep dts changes separate (which isn't a bad idea in itself). Regards, Bjorn
diff --git a/arch/arm64/boot/dts/qcom/sc7180-trogdor-lazor-r0.dts b/arch/arm64/boot/dts/qcom/sc7180-trogdor-lazor-r0.dts index 30e3e769d2b4..b69590a17cab 100644 --- a/arch/arm64/boot/dts/qcom/sc7180-trogdor-lazor-r0.dts +++ b/arch/arm64/boot/dts/qcom/sc7180-trogdor-lazor-r0.dts @@ -14,17 +14,6 @@ / { compatible = "google,lazor-rev0", "qcom,sc7180"; }; -&pp3300_hub { - /* pp3300_l7c is used to power the USB hub */ - /delete-property/regulator-always-on; - /delete-property/regulator-boot-on; -}; - -&pp3300_l7c { - regulator-always-on; - regulator-boot-on; -}; - &sn65dsi86_out { /* * Lane 0 was incorrectly mapped on the cable, but we've now decided @@ -33,3 +22,11 @@ &sn65dsi86_out { */ lane-polarities = <1 0>; }; + +&usb_hub_2_x { + vdd-supply = <&pp3300_l7c>; +}; + +&usb_hub_3_x { + vdd-supply = <&pp3300_l7c>; +}; diff --git a/arch/arm64/boot/dts/qcom/sc7180-trogdor-lazor-r1.dts b/arch/arm64/boot/dts/qcom/sc7180-trogdor-lazor-r1.dts index c2ef06367baf..25f178c3c6eb 100644 --- a/arch/arm64/boot/dts/qcom/sc7180-trogdor-lazor-r1.dts +++ b/arch/arm64/boot/dts/qcom/sc7180-trogdor-lazor-r1.dts @@ -14,13 +14,11 @@ / { compatible = "google,lazor-rev1", "google,lazor-rev2", "qcom,sc7180"; }; -&pp3300_hub { - /* pp3300_l7c is used to power the USB hub */ - /delete-property/regulator-always-on; - /delete-property/regulator-boot-on; + +&usb_hub_2_x { + vdd-supply = <&pp3300_l7c>; }; -&pp3300_l7c { - regulator-always-on; - regulator-boot-on; +&usb_hub_3_x { + vdd-supply = <&pp3300_l7c>; }; diff --git a/arch/arm64/boot/dts/qcom/sc7180-trogdor-pompom-r1.dts b/arch/arm64/boot/dts/qcom/sc7180-trogdor-pompom-r1.dts index 76a130bad60a..8467ff41e6d5 100644 --- a/arch/arm64/boot/dts/qcom/sc7180-trogdor-pompom-r1.dts +++ b/arch/arm64/boot/dts/qcom/sc7180-trogdor-pompom-r1.dts @@ -34,13 +34,10 @@ &pm6150_adc_tm { /delete-node/ charger-thermistor@0; }; -&pp3300_hub { - /* pp3300_l7c is used to power the USB hub */ - /delete-property/regulator-always-on; - /delete-property/regulator-boot-on; +&usb_hub_2_x { + vdd-supply = <&pp3300_l7c>; }; -&pp3300_l7c { - regulator-always-on; - regulator-boot-on; +&usb_hub_3_x { + vdd-supply = <&pp3300_l7c>; }; diff --git a/arch/arm64/boot/dts/qcom/sc7180-trogdor-r1.dts b/arch/arm64/boot/dts/qcom/sc7180-trogdor-r1.dts index 457c25499863..0cbb7a68d58b 100644 --- a/arch/arm64/boot/dts/qcom/sc7180-trogdor-r1.dts +++ b/arch/arm64/boot/dts/qcom/sc7180-trogdor-r1.dts @@ -43,17 +43,6 @@ &panel { compatible = "auo,b116xa01"; }; -&pp3300_hub { - /* pp3300_l7c is used to power the USB hub */ - /delete-property/regulator-always-on; - /delete-property/regulator-boot-on; -}; - -&pp3300_l7c { - regulator-always-on; - regulator-boot-on; -}; - &sdhc_2 { status = "okay"; }; @@ -62,6 +51,14 @@ &trackpad { interrupts = <58 IRQ_TYPE_EDGE_FALLING>; }; +&usb_hub_2_x { + vdd-supply = <&pp3300_l7c>; +}; + +&usb_hub_3_x { + vdd-supply = <&pp3300_l7c>; +}; + /* PINCTRL - modifications to sc7180-trogdor.dtsi */ &trackpad_int_1v8_odl { diff --git a/arch/arm64/boot/dts/qcom/sc7180-trogdor.dtsi b/arch/arm64/boot/dts/qcom/sc7180-trogdor.dtsi index d4f4441179fc..be06938872ca 100644 --- a/arch/arm64/boot/dts/qcom/sc7180-trogdor.dtsi +++ b/arch/arm64/boot/dts/qcom/sc7180-trogdor.dtsi @@ -202,7 +202,7 @@ pp3300_hub: pp3300-hub { pinctrl-names = "default"; pinctrl-0 = <&en_pp3300_hub>; - regulator-always-on; + /* The BIOS leaves this regulator on */ regulator-boot-on; vin-supply = <&pp3300_a>; @@ -839,6 +839,24 @@ &usb_1 { &usb_1_dwc3 { dr_mode = "host"; + #address-cells = <1>; + #size-cells = <0>; + + /* 2.x hub on port 1 */ + usb_hub_2_x: hub@1 { + compatible = "usbbda,5411"; + reg = <1>; + vdd-supply = <&pp3300_hub>; + companion-hub = <&usb_hub_3_x>; + }; + + /* 3.x hub on port 2 */ + usb_hub_3_x: hub@2 { + compatible = "usbbda,411"; + reg = <2>; + vdd-supply = <&pp3300_hub>; + companion-hub = <&usb_hub_2_x>; + }; }; &usb_1_hsphy {