On 7/13/23 19:14, Tim Harvey wrote:

On Fri, May 19, 2023 at 11:59 AM Tim Harvey tharvey@gateworks.com wrote:

...

Remove the regulator_set_enable() call from device probe which resolves a regulator imbalance. This is unnecessary as regulator_set_enable() will be called when ehci_register calls the init_after_reset hook.

Suggested-by: Marek Vasut marex@denx.de Signed-off-by: Tim Harvey tharvey@gateworks.com

---

drivers/usb/host/ehci-mx6.c | 16 ---------------- 1 file changed, 16 deletions(-)

diff --git a/drivers/usb/host/ehci-mx6.c b/drivers/usb/host/ehci-mx6.c index fae20838c60a..a9ed5e7a0d53 100644 --- a/drivers/usb/host/ehci-mx6.c +++ b/drivers/usb/host/ehci-mx6.c @@ -709,18 +709,6 @@ static int ehci_usb_probe(struct udevice *dev) goto err_regulator; #endif

-#if CONFIG_IS_ENABLED(DM_REGULATOR)

•   if (priv->vbus_supply) {
  

•           ret = regulator_set_enable(priv->vbus_supply,
  

•                                      (type == USB_INIT_DEVICE) ?
  

•                                      false : true);
  

•           if (ret && ret != -ENOSYS) {
  

•                   printf("Error enabling VBUS supply (ret=%i)\n", ret);
  

•                   goto err_clk;
  

•           }
  

•   }
  

-#endif

•    if (priv->init_type == USB_INIT_HOST) {
             setbits_le32(&ehci->usbmode, CM_HOST);
             writel(mx6_portsc(priv->phy_type), &ehci->portsc);
  

@@ -744,10 +732,6 @@ err_phy: generic_shutdown_phy(&priv->phy); err_regulator: #endif -#if CONFIG_IS_ENABLED(DM_REGULATOR)

•   if (priv->vbus_supply)
  

•           regulator_set_enable(priv->vbus_supply, false);
  

-#endif err_clk: #if CONFIG_IS_ENABLED(CLK) clk_disable(&priv->clk); -- 2.25.1

Hi Marek,

I think this goes in your tree. It got delegated to Stefano in patchwork but I'm not sure why as you maintain USB.

Applied to usb/master, thanks for the reminder (I really appreciate that)