On 31/03/20, Peter Chen wrote:
>
> > > compatible = "fsl,imx8qxp-lpcg";
> > > reg = <0x5b280000 0x10000>;
> > > #clock-cells = <1>;
> > > bit-offset = <0 4 16 20 24 28>;
> > > clocks = <&clk IMX_SC_R_USB_2 IMX_SC_PM_CLK_PER>,
> > > <&clk IMX_SC_R_USB_2 IMX_SC_PM_CLK_MISC>,
> > > <&conn_ipg_clk>,
> > > <&conn_ipg_clk>,
> > > <&conn_ipg_clk>,
> > > <&clk IMX_SC_R_USB_2 IMX_SC_PM_CLK_MST_BUS>;
> > > clock-output-names = "usb3_app_clk",
> > > "usb3_lpm_clk",
> > > "usb3_ipg_clk",
> > > "usb3_core_pclk",
> > > "usb3_phy_clk",
> > > "usb3_aclk";
> > > power-domains = <&pd IMX_SC_R_USB_2_PHY>;
> > > };
> > >
> > > usbotg3: usb3@5b110000 {
> > > compatible = "fsl,imx8qm-usb3";
> > > #address-cells = <1>;
> > > #size-cells = <1>;
> > > ranges;
> > > reg = <0x5B110000 0x10000>;
> > > clocks = <&usb3_lpcg 1>,
> > > <&usb3_lpcg 0>,
> > > <&usb3_lpcg 5>,
> > > <&usb3_lpcg 2>,
> > > <&usb3_lpcg 3>;
> > > clock-names = "usb3_lpm_clk", "usb3_bus_clk", "usb3_aclk",
> > > "usb3_ipg_clk", "usb3_core_pclk";
> > > assigned-clocks = <&clk IMX_SC_R_USB_2
> > IMX_SC_PM_CLK_PER>,
> > > <&clk IMX_SC_R_USB_2 IMX_SC_PM_CLK_MISC>,
> > > <&clk IMX_SC_R_USB_2 IMX_SC_PM_CLK_MST_BUS>;
> > > assigned-clock-rates = <125000000>, <12000000>, <250000000>;
> > > power-domains = <&pd IMX_SC_R_USB_2>;
> > > status = "disabled";
> > >
> > > usbotg3_cdns3: cdns3 {
> > > compatible = "cdns,usb3";
> > > #address-cells = <1>;
> > > #size-cells = <1>;
> > > interrupt-parent = <&gic>;
> > > interrupts = <GIC_SPI 271 IRQ_TYPE_LEVEL_HIGH>,
> > > <GIC_SPI 271 IRQ_TYPE_LEVEL_HIGH>,
> > > <GIC_SPI 271 IRQ_TYPE_LEVEL_HIGH>;
> > > interrupt-names = "host", "peripheral", "otg";
> > > reg = <0x5B130000 0x10000>, /* memory area for HOST
> > registers */
> > > <0x5B140000 0x10000>, /* memory area for
> > DEVICE registers */
> > > <0x5B120000 0x10000>; /* memory area for
> > OTG/DRD registers */
> > > reg-names = "xhci", "dev", "otg";
> > > phys = <&usb3_phy>;
> > > phy-names = "cdns3,usb3-phy";
> > > status = "disabled";
> > > };
> > > };
> >
> > Hello Peter,
> >
> > thx for the dts example. Now I get "wait clkvld timeout" in "cdns_imx_noncore_init()"
> > So USB3_SSPHY_STATUS register seems to be wrong.
> >
> > dmesg | grep imx
> > [ 1.065445] cdns3-imx 5b110000.usb3: Adding to iommu group 0
> > [ 1.257309] imx-scu scu: NXP i.MX SCU Initialized
> > [ 1.275489] imx-scu-clk: probe of gpt0_clk failed with error -5
> > [ 1.291143] imx-scu-clk: probe of pwm_clk failed with error -22
> > [ 1.302517] imx-scu-clk: probe of lcd_clk failed with error -22
> > [ 1.341405] imx8qm-pinctrl scu:pinctrl: initialized IMX pinctrl driver
> > [ 3.029484] cdns3-imx 5b110000.usb3: wait clkvld timeout
> > [ 3.034891] cdns3-imx: probe of 5b110000.usb3 failed with error -110
> > [ 3.237393] sdhci-esdhc-imx 5b030000.mmc: Got CD GPIO
> > [ 3.242468] sdhci-esdhc-imx 5b030000.mmc: Got WP GPIO
> > [ 3.316687] imx8qxp-lpcg-clk 5a4a0000.clock-controller: ignoring dependency for
> > device, assuming no driver
> >
>
> Make sure the five clocks in dts are correct, and print the USB3_SSPHY_STATUS if
> timeout still exists.
ok thx, we got it fixed by replacing:
clocks = <&usb3_lpcg 1>,
<&usb3_lpcg 0>,
<&usb3_lpcg 4>,
<&usb3_lpcg 2>,
<&usb3_lpcg 3>;
with this:
clocks = <&usb3_lpcg IMX_LPCG_CLK_1>,
<&usb3_lpcg IMX_LPCG_CLK_0>,
<&usb3_lpcg IMX_LPCG_CLK_7>,
<&usb3_lpcg IMX_LPCG_CLK_4>,
<&usb3_lpcg IMX_LPCG_CLK_5>;
now the "wait clkvld timeout" is gone and the USB3_SSPHY_STATUS register
is very similar to another imx8qm device which runs linux-imx from NXP.
Now I try to get into USB HOST mode:
--- a/drivers/usb/cdns3/cdns3-imx.c
+++ b/drivers/usb/cdns3/cdns3-imx.c
@@ -113,11 +115,11 @@ static int cdns_imx_noncore_init(struct cdns_imx *data)
udelay(1);
value = cdns_imx_readl(data, USB3_CORE_CTRL1);
- value = (value & ~MODE_STRAP_MASK) | OTG_MODE | OC_DISABLE;
+ value = (value & ~MODE_STRAP_MASK) | HOST_MODE | OC_DISABLE;
cdns_imx_writel(data, USB3_CORE_CTRL1, value);
Is this change necessary?
Do I need special changes on xhci.c do get it work with cdns-imx in host mode?
I'am stuck at:
[ 16.697525] xhci-hcd xhci-hcd.0.auto: can't setup: -110
I already tried to add the Intel quirk:
udelay(1000) in xhci_reset() but without success.
Is the host mode working on your side?
Best regards,
Oliver
