| Message ID | 20190506234631.113226-2-chris.brandt@renesas.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | usb: Add host and device support for RZ/A2 | expand |
Hi Chris, On Tue, May 7, 2019 at 1:47 AM Chris Brandt <chris.brandt@renesas.com> wrote: > The RZ/A2 has optional dedicated 48MHz crystal inputs, which adds a new > register setting when used. > > Signed-off-by: Chris Brandt <chris.brandt@renesas.com> Thanks for your patch! > --- a/drivers/phy/renesas/phy-rcar-gen3-usb2.c > +++ b/drivers/phy/renesas/phy-rcar-gen3-usb2.c > @@ -630,6 +635,9 @@ static int rcar_gen3_phy_usb2_probe(struct platform_device *pdev) > } > } > > + if (of_property_read_bool(dev->of_node, "renesas,uses_usb_x1")) > + channel->uses_usb_x1 = true; > + Perhaps this can be checked some other way (e.g. by checking for a non-zero clock rate of the USB_X1 clock referenced from DT), thus removing the need for adding a custom property? Gr{oetje,eeting}s, Geert
Hi Geert, On Tue, May 07, 2019, Geert Uytterhoeven wrote: > > + if (of_property_read_bool(dev->of_node, "renesas,uses_usb_x1")) > > + channel->uses_usb_x1 = true; > > + > > Perhaps this can be checked some other way (e.g. by checking for a non- > zero > clock rate of the USB_X1 clock referenced from DT), thus removing the need > for > adding a custom property? Good point. I've done that for other drivers before and it worked well. I'll take this out...one less property to set :) Question: Since the driver depends on this, should I mention this in the dt-bindings documentation ? Thanks, Chris
Hi Chris, On Tue, May 7, 2019 at 1:00 PM Chris Brandt <Chris.Brandt@renesas.com> wrote: > On Tue, May 07, 2019, Geert Uytterhoeven wrote: > > > + if (of_property_read_bool(dev->of_node, "renesas,uses_usb_x1")) > > > + channel->uses_usb_x1 = true; > > > + > > > > Perhaps this can be checked some other way (e.g. by checking for a non- > > zero > > clock rate of the USB_X1 clock referenced from DT), thus removing the need > > for > > adding a custom property? > > Good point. I've done that for other drivers before and it worked well. > > I'll take this out...one less property to set :) > > Question: Since the driver depends on this, should I mention this in the > dt-bindings documentation ? Yes, this should be described in the section about the clocks property. Gr{oetje,eeting}s, Geert
Hi Geert, On Tue, May 07, 2019, Geert Uytterhoeven wrote: > > + if (of_property_read_bool(dev->of_node, "renesas,uses_usb_x1")) > > + channel->uses_usb_x1 = true; > > + > > Perhaps this can be checked some other way (e.g. by checking for a non- > zero > clock rate of the USB_X1 clock referenced from DT), thus removing the need > for > adding a custom property? Currently, there is no USB_X1 in DT like there is for RZ/A1. For RZ/A2, those are dedicated pins that belong to the USB HW block itself. They do not feed into the system CPG or any dividers, so I never included it in the .dtsi. So with that said, does a uses-usb-x1 property make more sense? Chris
Hi Chris, On Tue, May 7, 2019 at 5:43 PM Chris Brandt <Chris.Brandt@renesas.com> wrote: > On Tue, May 07, 2019, Geert Uytterhoeven wrote: > > > + if (of_property_read_bool(dev->of_node, "renesas,uses_usb_x1")) > > > + channel->uses_usb_x1 = true; > > > + > > > > Perhaps this can be checked some other way (e.g. by checking for a non- > > zero > > clock rate of the USB_X1 clock referenced from DT), thus removing the need > > for > > adding a custom property? > > Currently, there is no USB_X1 in DT like there is for RZ/A1. > > For RZ/A2, those are dedicated pins that belong to the USB HW block > itself. They do not feed into the system CPG or any dividers, so I > never included it in the .dtsi. Like pcie_bus_clk on R-Car? We do have that in DT, with a "clock" link to it from the PCIe device node. After all, it is provided by an external clock crystal, and consumed by the PCIe device. > So with that said, does a uses-usb-x1 property make more sense? No ;-) Gr{oetje,eeting}s, Geert
Hi Geert, On Tue, May 07, 2019 1, Geert Uytterhoeven wrote: > > So with that said, does a uses-usb-x1 property make more sense? > > No ;-) So.... I guess the first patch in the series needs to add this to the .dtsi: usb_x1_clk: usb_x1 { #clock-cells = <0>; compatible = "fixed-clock"; /* If clk present, value must be set by board */ clock-frequency = <0>; }; Then I can reference "usb_x1" in the driver and see if it is set to non-zero. What do you think? Chris
Hi Chris, On Tue, May 7, 2019 at 10:27 PM Chris Brandt <Chris.Brandt@renesas.com> wrote: > On Tue, May 07, 2019 1, Geert Uytterhoeven wrote: > > > So with that said, does a uses-usb-x1 property make more sense? > > > > No ;-) > > So.... > > I guess the first patch in the series needs to add this to the .dtsi: > > usb_x1_clk: usb_x1 { > #clock-cells = <0>; > compatible = "fixed-clock"; > /* If clk present, value must be set by board */ > clock-frequency = <0>; > }; > > Then I can reference "usb_x1" in the driver and see if it is set to > non-zero. > > What do you think? Exactly! Gr{oetje,eeting}s, Geert
diff --git a/drivers/phy/renesas/phy-rcar-gen3-usb2.c b/drivers/phy/renesas/phy-rcar-gen3-usb2.c index 1322185a00a2..218b32e458cb 100644 --- a/drivers/phy/renesas/phy-rcar-gen3-usb2.c +++ b/drivers/phy/renesas/phy-rcar-gen3-usb2.c @@ -34,6 +34,7 @@ #define USB2_VBCTRL 0x60c #define USB2_LINECTRL1 0x610 #define USB2_ADPCTRL 0x630 +#define USB2_PHYCLK_CTRL 0x644 /* INT_ENABLE */ #define USB2_INT_ENABLE_UCOM_INTEN BIT(3) @@ -110,6 +111,7 @@ struct rcar_gen3_chan { bool extcon_host; bool is_otg_channel; bool uses_otg_pins; + bool uses_usb_x1; }; /* @@ -391,6 +393,9 @@ static int rcar_gen3_phy_usb2_init(struct phy *p) void __iomem *usb2_base = channel->base; u32 val; + if (channel->uses_usb_x1) + writel(0x00000001, usb2_base + USB2_PHYCLK_CTRL); + /* Initialize USB2 part */ val = readl(usb2_base + USB2_INT_ENABLE); val |= USB2_INT_ENABLE_UCOM_INTEN | rphy->int_enable_bits; @@ -630,6 +635,9 @@ static int rcar_gen3_phy_usb2_probe(struct platform_device *pdev) } } + if (of_property_read_bool(dev->of_node, "renesas,uses_usb_x1")) + channel->uses_usb_x1 = true; + /* * devm_phy_create() will call pm_runtime_enable(&phy->dev); * And then, phy-core will manage runtime pm for this device.
The RZ/A2 has optional dedicated 48MHz crystal inputs, which adds a new register setting when used. Signed-off-by: Chris Brandt <chris.brandt@renesas.com> --- drivers/phy/renesas/phy-rcar-gen3-usb2.c | 8 ++++++++ 1 file changed, 8 insertions(+)