[U-Boot] [PATCH 1/2] usb: host: xhci: make sure to power up PHY
some boards won't work if the PHY isn't explicitly powered up.
Signed-off-by: Felipe Balbi balbi@ti.com --- drivers/usb/host/xhci-omap.c | 1 + 1 file changed, 1 insertion(+)
diff --git a/drivers/usb/host/xhci-omap.c b/drivers/usb/host/xhci-omap.c index e667810..912b2bd 100644 --- a/drivers/usb/host/xhci-omap.c +++ b/drivers/usb/host/xhci-omap.c @@ -98,6 +98,7 @@ static int omap_xhci_core_init(struct omap_xhci *omap) { int ret = 0;
+ usb_phy_power(1); omap_enable_phy(omap);
ret = dwc3_core_init(omap->dwc3_reg);
Newer AM437x silicon requires us to explicitly power up the USB2 PHY. By implementing usb_phy_power() we can achieve that.
Signed-off-by: Felipe Balbi balbi@ti.com --- arch/arm/include/asm/arch-am33xx/hardware_am43xx.h | 5 +++++ drivers/usb/phy/omap_usb_phy.c | 11 ++++++++++- 2 files changed, 15 insertions(+), 1 deletion(-)
diff --git a/arch/arm/include/asm/arch-am33xx/hardware_am43xx.h b/arch/arm/include/asm/arch-am33xx/hardware_am43xx.h index b470319..efdecf4 100644 --- a/arch/arm/include/asm/arch-am33xx/hardware_am43xx.h +++ b/arch/arm/include/asm/arch-am33xx/hardware_am43xx.h @@ -43,6 +43,11 @@ #define VTP0_CTRL_ADDR 0x44E10E0C #define VTP1_CTRL_ADDR 0x48140E10
+/* USB CTRL Base Address */ +#define USB1_CTRL 0x44e10628 +#define USB1_CTRL_CM_PWRDN BIT(0) +#define USB1_CTRL_OTG_PWRDN BIT(1) + /* DDR Base address */ #define DDR_PHY_CMD_ADDR 0x44E12000 #define DDR_PHY_DATA_ADDR 0x44E120C8 diff --git a/drivers/usb/phy/omap_usb_phy.c b/drivers/usb/phy/omap_usb_phy.c index af46db2..53778b3 100644 --- a/drivers/usb/phy/omap_usb_phy.c +++ b/drivers/usb/phy/omap_usb_phy.c @@ -222,7 +222,16 @@ static void am437x_enable_usb2_phy2(struct omap_xhci *omap)
void usb_phy_power(int on) { - return; + /* + * In order to get USB working on newer AM43xx silicon, we must + * explicitly power the USB PHYs. + */ + if (on) + clrbits_le32(USB1_CTRL, USB1_CTRL_CM_PWRDN | + USB1_CTRL_OTG_PWRDN); + else + setbits_le32(USB1_CTRL, USB1_CTRL_CM_PWRDN | + USB1_CTRL_OTG_PWRDN); } #endif /* CONFIG_AM437X_USB2PHY2_HOST */
On Wed, Jul 09, 2014 at 11:02:43PM -0500, Felipe Balbi wrote:
Newer AM437x silicon requires us to explicitly power up the USB2 PHY. By implementing usb_phy_power() we can achieve that.
Signed-off-by: Felipe Balbi balbi@ti.com
the only change here is the use of {set,clr}bits_le32()
arch/arm/include/asm/arch-am33xx/hardware_am43xx.h | 5 +++++ drivers/usb/phy/omap_usb_phy.c | 11 ++++++++++- 2 files changed, 15 insertions(+), 1 deletion(-)
diff --git a/arch/arm/include/asm/arch-am33xx/hardware_am43xx.h b/arch/arm/include/asm/arch-am33xx/hardware_am43xx.h index b470319..efdecf4 100644 --- a/arch/arm/include/asm/arch-am33xx/hardware_am43xx.h +++ b/arch/arm/include/asm/arch-am33xx/hardware_am43xx.h @@ -43,6 +43,11 @@ #define VTP0_CTRL_ADDR 0x44E10E0C #define VTP1_CTRL_ADDR 0x48140E10
+/* USB CTRL Base Address */ +#define USB1_CTRL 0x44e10628 +#define USB1_CTRL_CM_PWRDN BIT(0) +#define USB1_CTRL_OTG_PWRDN BIT(1)
/* DDR Base address */ #define DDR_PHY_CMD_ADDR 0x44E12000 #define DDR_PHY_DATA_ADDR 0x44E120C8 diff --git a/drivers/usb/phy/omap_usb_phy.c b/drivers/usb/phy/omap_usb_phy.c index af46db2..53778b3 100644 --- a/drivers/usb/phy/omap_usb_phy.c +++ b/drivers/usb/phy/omap_usb_phy.c @@ -222,7 +222,16 @@ static void am437x_enable_usb2_phy2(struct omap_xhci *omap)
void usb_phy_power(int on) {
- return;
- /*
* In order to get USB working on newer AM43xx silicon, we must* explicitly power the USB PHYs.*/- if (on)
clrbits_le32(USB1_CTRL, USB1_CTRL_CM_PWRDN |USB1_CTRL_OTG_PWRDN);- else
setbits_le32(USB1_CTRL, USB1_CTRL_CM_PWRDN |USB1_CTRL_OTG_PWRDN);} #endif /* CONFIG_AM437X_USB2PHY2_HOST */
-- 2.0.0.390.gcb682f8
On Thursday, July 10, 2014 at 06:06:06 AM, Felipe Balbi wrote:
On Wed, Jul 09, 2014 at 11:02:43PM -0500, Felipe Balbi wrote:
Newer AM437x silicon requires us to explicitly power up the USB2 PHY. By implementing usb_phy_power() we can achieve that.
Signed-off-by: Felipe Balbi balbi@ti.com
the only change here is the use of {set,clr}bits_le32()
Well, given the pressure before release, I picked the previous one (without clrsetbits). I think the original will also generate less code, so let's keep it at that. Sorry I've been pushing you around.
Best regards, Marek Vasut
participants (2)
-
Felipe Balbi -
Marek Vasut